Search

Joint-embedding predictive architectures (JEPAs) have shown substantial promise in self-supervised representation learning, yet their application in generative modeling remains underexplored. Conversely, diffusion models have demonstrated significant efficacy in modeling arbitrary probability distributions. In this paper, we introduce Denoising with a Joint-Embedding Predictive Architecture (D-JEPA), pioneering the integration of JEPA within generative modeling. By recognizing JEPA as a form of masked image modeling, we reinterpret it as a generalized next-token prediction strategy, facilitating data generation in an auto-regressive manner. Furthermore, we incorporate diffusion loss to model the per-token probability distribution, enabling data generation in a continuous space. We also adapt flow matching loss as an alternative to diffusion loss, thereby enhancing the flexibility of D-JEPA. Empirically, with increased GFLOPs, D-JEPA consistently achieves lower FID scores with fewer training epochs, indicating its good scalability. Our base, large, and huge models outperform all previous generative models across all scales on class-conditional ImageNet benchmarks. Beyond image generation, D-JEPA is well-suited for other continuous data modeling, including video and audio., Comment: 38 pages

Introducing user-specified visual concepts in image editing is highly practical as these concepts convey the user's intent more precisely than text-based descriptions. We propose FreeEdit, a novel approach for achieving such reference-based image editing, which can accurately reproduce the visual concept from the reference image based on user-friendly language instructions. Our approach leverages the multi-modal instruction encoder to encode language instructions to guide the editing process. This implicit way of locating the editing area eliminates the need for manual editing masks. To enhance the reconstruction of reference details, we introduce the Decoupled Residual ReferAttention (DRRA) module. This module is designed to integrate fine-grained reference features extracted by a detail extractor into the image editing process in a residual way without interfering with the original self-attention. Given that existing datasets are unsuitable for reference-based image editing tasks, particularly due to the difficulty in constructing image triplets that include a reference image, we curate a high-quality dataset, FreeBench, using a newly developed twice-repainting scheme. FreeBench comprises the images before and after editing, detailed editing instructions, as well as a reference image that maintains the identity of the edited object, encompassing tasks such as object addition, replacement, and deletion. By conducting phased training on FreeBench followed by quality tuning, FreeEdit achieves high-quality zero-shot editing through convenient language instructions. We conduct extensive experiments to evaluate the effectiveness of FreeEdit across multiple task types, demonstrating its superiority over existing methods. The code will be available at: https://freeedit.github.io/., Comment: 14 pages, 14 figures, project website: https://freeedit.github.io/

Panoptic narrative grounding (PNG), whose core target is fine-grained image-text alignment, requires a panoptic segmentation of referred objects given a narrative caption. Previous discriminative methods achieve only weak or coarse-grained alignment by panoptic segmentation pretraining or CLIP model adaptation. Given the recent progress of text-to-image Diffusion models, several works have shown their capability to achieve fine-grained image-text alignment through cross-attention maps and improved general segmentation performance. However, the direct use of phrase features as static prompts to apply frozen Diffusion models to the PNG task still suffers from a large task gap and insufficient vision-language interaction, yielding inferior performance. Therefore, we propose an Extractive-Injective Phrase Adapter (EIPA) bypass within the Diffusion UNet to dynamically update phrase prompts with image features and inject the multimodal cues back, which leverages the fine-grained image-text alignment capability of Diffusion models more sufficiently. In addition, we also design a Multi-Level Mutual Aggregation (MLMA) module to reciprocally fuse multi-level image and phrase features for segmentation refinement. Extensive experiments on the PNG benchmark show that our method achieves new state-of-the-art performance., Comment: Accepted by ACM MM 2024

Rapid advancements have been made in extending Large Language Models (LLMs) to Large Multi-modal Models (LMMs). However, extending input modality of LLMs to video data remains a challenging endeavor, especially for long videos. Due to insufficient access to large-scale high-quality video data and the excessive compression of visual features, current methods exhibit limitations in effectively processing long videos. In this paper, we introduce Kangaroo, a powerful Video LMM aimed at addressing these challenges. Confronted with issue of inadequate training data, we develop a data curation system to build a large-scale dataset with high-quality annotations for vision-language pre-training and instruction tuning. In addition, we design a curriculum training pipeline with gradually increasing resolution and number of input frames to accommodate long videos. Evaluation results demonstrate that, with 8B parameters, Kangaroo achieves state-of-the-art performance across a variety of video understanding benchmarks while exhibiting competitive results on others. Particularly, on benchmarks specialized for long videos, Kangaroo excels some larger models with over 10B parameters and proprietary models.

In this paper, we propose an Audio-Language-Referenced SAM 2 (AL-Ref-SAM 2) pipeline to explore the training-free paradigm for audio and language-referenced video object segmentation, namely AVS and RVOS tasks. The intuitive solution leverages GroundingDINO to identify the target object from a single frame and SAM 2 to segment the identified object throughout the video, which is less robust to spatiotemporal variations due to a lack of video context exploration. Thus, in our AL-Ref-SAM 2 pipeline, we propose a novel GPT-assisted Pivot Selection (GPT-PS) module to instruct GPT-4 to perform two-step temporal-spatial reasoning for sequentially selecting pivot frames and pivot boxes, thereby providing SAM 2 with a high-quality initial object prompt. Within GPT-PS, two task-specific Chain-of-Thought prompts are designed to unleash GPT's temporal-spatial reasoning capacity by guiding GPT to make selections based on a comprehensive understanding of video and reference information. Furthermore, we propose a Language-Binded Reference Unification (LBRU) module to convert audio signals into language-formatted references, thereby unifying the formats of AVS and RVOS tasks in the same pipeline. Extensive experiments on both tasks show that our training-free AL-Ref-SAM 2 pipeline achieves performances comparable to or even better than fully-supervised fine-tuning methods. The code is available at: https://github.com/appletea233/AL-Ref-SAM2.

Incorporating a temporal dimension into pretrained image diffusion models for video generation is a prevalent approach. However, this method is computationally demanding and necessitates large-scale video datasets. More critically, the heterogeneity between image and video datasets often results in catastrophic forgetting of the image expertise. Recent attempts to directly extract video snippets from image diffusion models have somewhat mitigated these problems. Nevertheless, these methods can only generate brief video clips with simple movements and fail to capture fine-grained motion or non-grid deformation. In this paper, we propose a novel Zero-Shot video Sampling algorithm, denoted as $\mathcal{ZS}^2$, capable of directly sampling high-quality video clips from existing image synthesis methods, such as Stable Diffusion, without any training or optimization. Specifically, $\mathcal{ZS}^2$ utilizes the dependency noise model and temporal momentum attention to ensure content consistency and animation coherence, respectively. This ability enables it to excel in related tasks, such as conditional and context-specialized video generation and instruction-guided video editing. Experimental results demonstrate that $\mathcal{ZS}^2$ achieves state-of-the-art performance in zero-shot video generation, occasionally outperforming recent supervised methods. Homepage: \url{https://densechen.github.io/zss/}.

The Gaussian diffusion model, initially designed for image generation, has recently been adapted for 3D point cloud generation. However, these adaptations have not fully considered the intrinsic geometric characteristics of 3D shapes, thereby constraining the diffusion model's potential for 3D shape manipulation. To address this limitation, we introduce a novel deformable 3D shape diffusion model that facilitates comprehensive 3D shape manipulation, including point cloud generation, mesh deformation, and facial animation. Our approach innovatively incorporates a differential deformation kernel, which deconstructs the generation of geometric structures into successive non-rigid deformation stages. By leveraging a probabilistic diffusion model to simulate this step-by-step process, our method provides a versatile and efficient solution for a wide range of applications, spanning from graphics rendering to facial expression animation. Empirical evidence highlights the effectiveness of our approach, demonstrating state-of-the-art performance in point cloud generation and competitive results in mesh deformation. Additionally, extensive visual demonstrations reveal the significant potential of our approach for practical applications. Our method presents a unique pathway for advancing 3D shape manipulation and unlocking new opportunities in the realm of virtual reality.

Three-dimensional (3D) dissipative soliton existed in spatiotemporal mode-locked (STML) multimode fiber laser has been demonstrated to be a promising formalism for generating high-energy femtosecond pulses, which unfortunately exhibit diverse spatiotemporal dynamics that have not been fully understood. Completely modeling the STML multimode fiber lasers can shed new light on the underlying physics of the spatiotemporal dynamics and thus better manipulate the generation of high-quality energic femtosecond pulses, which however is still largely unmet. To this end, here we theoretically investigate a dynamic multimode gain model of the STML multimode fiber laser by exploring the multimode rate equation (MMRE) in the framework of generalized multimode nonlinear Schr\"odinger equation. Using this dynamic multimode gain model, the attractor dissection theory is revisited to understand the dominant effects that determine the modal composition of 3D dissipative soliton. Specifically, by varying the numerical aperture of the multimode gain fiber (MMGF), different gain dynamics that correspond to distinct types of gain attractors are observed. As a result, two distinguishing STML operation regimes, respectively governed by the multimode gain effect and spatiotemporal saturable absorption, are identified. In the latter regime, especially, 3D dissipative solitons present bistability that there exist bifurcated solutions with two different linearly polarized (LP) mode compositions. To verify the theoretical findings, the experimental implementation shows that the state of STML can be switched between different LP modes, and confirms the presence of bistability. Particularly, the 3D-soliton shaping mechanism that is governed by the multimode gain effect is testified for the first time, to the best of our knowledge.

Dual-comb spectroscopy (DCS) with few-GHz tooth spacing that provides the optimal trade-off between spectral resolution and refresh rate is a powerful tool for measuring and analyzing rapidly evolving transient events. Despite such an exciting opportunity, existing technologies compromise either the spectral resolution or refresh rate, leaving few-GHz DCS with robust design largely unmet for frontier applications. In this work, we demonstrate a novel GHz DCS by exploring the multimode interference-mediated spectral filtering effect in an all-fiber ultrashort cavity configuration. The GHz single-cavity all-fiber dual-comb source is seeded by a dual-wavelength mode-locked fiber laser operating at fundamental repetition rates of about 1.0 GHz differing by 148 kHz, which has an excellent stability in the free-running state that the Allan deviation is only 101.7 mHz for an average time of 1 second. Thanks to the large repetition rate difference between the asynchronous dichromatic pulse trains, the GHz DCS enables a refresh time as short as 6.75 us, making it promising for studying nonrepeatable transient phenomena in real time. To this end, the practicality of the present GHz DCS is validated by successfully capturing the 'shock waves' of balloon and firecracker explosions outdoors. This GHz single-cavity all-fiber dual-comb system promises a noteworthy improvement in acquisition speed and reliability without sacrificing measurement accuracy, anticipated as a practical tool for high-speed applications.

Data mixing methods play a crucial role in semi-supervised learning (SSL), but their application is unexplored in long-tailed semi-supervised learning (LTSSL). The primary reason is that the in-batch mixing manner fails to address class imbalance. Furthermore, existing LTSSL methods mainly focus on re-balancing data quantity but ignore class-wise uncertainty, which is also vital for class balance. For instance, some classes with sufficient samples might still exhibit high uncertainty due to indistinguishable features. To this end, this paper introduces the Balanced and Entropy-based Mix (BEM), a pioneering mixing approach to re-balance the class distribution of both data quantity and uncertainty. Specifically, we first propose a class balanced mix bank to store data of each class for mixing. This bank samples data based on the estimated quantity distribution, thus re-balancing data quantity. Then, we present an entropy-based learning approach to re-balance class-wise uncertainty, including entropy-based sampling strategy, entropy-based selection module, and entropy-based class balanced loss. Our BEM first leverages data mixing for improving LTSSL, and it can also serve as a complement to the existing re-balancing methods. Experimental results show that BEM significantly enhances various LTSSL frameworks and achieves state-of-the-art performances across multiple benchmarks., Comment: This paper is accepted to CVPR 2024. The supplementary material is included

In recent years, text-image joint pre-training techniques have shown promising results in various tasks. However, in Optical Character Recognition (OCR) tasks, aligning text instances with their corresponding text regions in images poses a challenge, as it requires effective alignment between text and OCR-Text (referring to the text in images as OCR-Text to distinguish from the text in natural language) rather than a holistic understanding of the overall image content. In this paper, we propose a new pre-training method called OCR-Text Destylization Modeling (ODM) that transfers diverse styles of text found in images to a uniform style based on the text prompt. With ODM, we achieve better alignment between text and OCR-Text and enable pre-trained models to adapt to the complex and diverse styles of scene text detection and spotting tasks. Additionally, we have designed a new labeling generation method specifically for ODM and combined it with our proposed Text-Controller module to address the challenge of annotation costs in OCR tasks, allowing a larger amount of unlabeled data to participate in pre-training. Extensive experiments on multiple public datasets demonstrate that our method significantly improves performance and outperforms current pre-training methods in scene text detection and spotting tasks. Code is available at https://github.com/PriNing/ODM., Comment: Accepted by CVPR2024

We demonstrate a compact watt-level all polarization-maintaining (PM) femtosecond fiber laser source at 1100 nm. The fiber laser source is seeded by an all PM fiber mode-locked laser employing a nonlinear amplifying loop mirror. The seed laser can generate stable pulses at a fundamental repetition rate of 40.71 MHz with a signal-to-noise rate of >100 dB and an integrated relative intensity noise of only ~0.061%. After two-stage external amplification and pulse compression, an output power of ~1.47 W (corresponding to a pulse energy of ~36.1 nJ) and a pulse duration of ~251 fs are obtained. The 1100 nm femtosecond fiber laser is then employed as the excitation light source for multicolor multi-photon fluorescence microscopy of Chinese hamster ovary (CHO) cells stably expressing red fluorescent proteins.

Panoptic narrative grounding (PNG) aims to segment things and stuff objects in an image described by noun phrases of a narrative caption. As a multimodal task, an essential aspect of PNG is the visual-linguistic interaction between image and caption. The previous two-stage method aggregates visual contexts from offline-generated mask proposals to phrase features, which tend to be noisy and fragmentary. The recent one-stage method aggregates only pixel contexts from image features to phrase features, which may incur semantic misalignment due to lacking object priors. To realize more comprehensive visual-linguistic interaction, we propose to enrich phrases with coupled pixel and object contexts by designing a Phrase-Pixel-Object Transformer Decoder (PPO-TD), where both fine-grained part details and coarse-grained entity clues are aggregated to phrase features. In addition, we also propose a PhraseObject Contrastive Loss (POCL) to pull closer the matched phrase-object pairs and push away unmatched ones for aggregating more precise object contexts from more phrase-relevant object tokens. Extensive experiments on the PNG benchmark show our method achieves new state-of-the-art performance with large margins., Comment: Accepted by IJCAI 2023. Since the PNG benchmark adopts a different data partition manner from ours, we update the experimental results on the things/stuff/singulars/plurals subsets based on the PNG's code

Zero-shot video recognition (ZSVR) is a task that aims to recognize video categories that have not been seen during the model training process. Recently, vision-language models (VLMs) pre-trained on large-scale image-text pairs have demonstrated impressive transferability for ZSVR. To make VLMs applicable to the video domain, existing methods often use an additional temporal learning module after the image-level encoder to learn the temporal relationships among video frames. Unfortunately, for video from unseen categories, we observe an abnormal phenomenon where the model that uses spatial-temporal feature performs much worse than the model that removes temporal learning module and uses only spatial feature. We conjecture that improper temporal modeling on video disrupts the spatial feature of the video. To verify our hypothesis, we propose Feature Factorization to retain the orthogonal temporal feature of the video and use interpolation to construct refined spatial-temporal feature. The model using appropriately refined spatial-temporal feature performs better than the one using only spatial feature, which verifies the effectiveness of the orthogonal temporal feature for the ZSVR task. Therefore, an Orthogonal Temporal Interpolation module is designed to learn a better refined spatial-temporal video feature during training. Additionally, a Matching Loss is introduced to improve the quality of the orthogonal temporal feature. We propose a model called OTI for ZSVR by employing orthogonal temporal interpolation and the matching loss based on VLMs. The ZSVR accuracies on popular video datasets (i.e., Kinetics-600, UCF101 and HMDB51) show that OTI outperforms the previous state-of-the-art method by a clear margin.

We present a hardware-efficient architecture of convolutional neural network, which has a repvgg-like architecture. Flops or parameters are traditional metrics to evaluate the efficiency of networks which are not sensitive to hardware including computing ability and memory bandwidth. Thus, how to design a neural network to efficiently use the computing ability and memory bandwidth of hardware is a critical problem. This paper proposes a method how to design hardware-aware neural network. Based on this method, we designed EfficientRep series convolutional networks, which are high-computation hardware(e.g. GPU) friendly and applied in YOLOv6 object detection framework. YOLOv6 has published YOLOv6N/YOLOv6S/YOLOv6M/YOLOv6L models in v1 and v2 versions.

External driving of passive, nonlinear optical resonators has emerged over the past decade as a novel route for the generation of ultrashort optical pulses and corresponding broadband frequency combs. Whilst the pulse formation dynamics in such systems differ dramatically from those manifesting themselves in conventional mode-locked lasers, the demarcation between the two traditionally distinct paradigms has recently begun to blur, with demonstrations of hybrid systems incorporating both external driving and active media shown to offer specific advantages. Here we explore a new pathway for ultrashort pulse generation at the interface of externally-driven passive resonators and lasers. By leveraging the nonlinear Raman gain inherent to fused silica, we achieve deterministic generation of low-noise dissipative solitons with durations well below 100 fs via phase-coherent pulsed driving of resonators made of standard, commercially-available optical fibre. We explore and explain the physics of the new dissipative Raman soliton states, identifying scaling laws that govern the pulses' characteristics and that allow output repetition rates to be scaled at will without influencing the soliton duration. The scheme explored in our work enables the shortest ever pulses generated in resonators (active or passive) made from a single commercially-available optical fibre, and it has the potential to be transferred into a chip-scale format by using existing dispersion-engineered silica microresonators., Comment: 15 pages, 7 figures

It is well believed that the higher uncertainty in a word of the caption, the more inter-correlated context information is required to determine it. However, current image captioning methods usually consider the generation of all words in a sentence sequentially and equally. In this paper, we propose an uncertainty-aware image captioning framework, which parallelly and iteratively operates insertion of discontinuous candidate words between existing words from easy to difficult until converged. We hypothesize that high-uncertainty words in a sentence need more prior information to make a correct decision and should be produced at a later stage. The resulting non-autoregressive hierarchy makes the caption generation explainable and intuitive. Specifically, we utilize an image-conditioned bag-of-word model to measure the word uncertainty and apply a dynamic programming algorithm to construct the training pairs. During inference, we devise an uncertainty-adaptive parallel beam search technique that yields an empirically logarithmic time complexity. Extensive experiments on the MS COCO benchmark reveal that our approach outperforms the strong baseline and related methods on both captioning quality as well as decoding speed., Comment: Accepted by AAAI2023

Ensemble of machine learning models yields improved performance as well as robustness. However, their memory requirements and inference costs can be prohibitively high. Knowledge distillation is an approach that allows a single model to efficiently capture the approximate performance of an ensemble while showing poor scalability as demand for re-training when introducing new teacher models. In this paper, we study if we can utilize the meta-learning strategy to directly predict the parameters of a single model with comparable performance of an ensemble. Hereto, we introduce WeightFormer, a Transformer-based model that can predict student network weights layer by layer in a forward pass, according to the teacher model parameters. The proprieties of WeightFormer are investigated on the CIFAR-10, CIFAR-100, and ImageNet datasets for model structures of VGGNet-11, ResNet-50, and ViT-B/32, where it demonstrates that our method can achieve approximate classification performance of an ensemble and outperforms both the single network and standard knowledge distillation. More encouragingly, we show that WeightFormer results can further exceeds average ensemble with minor fine-tuning. Importantly, our task along with the model and results can potentially lead to a new, more efficient, and scalable paradigm of ensemble networks parameter learning., Comment: technique report

Over past few years afterward the birth of ResNet, skip connection has become the defacto standard for the design of modern architectures due to its widespread adoption, easy optimization and proven performance. Prior work has explained the effectiveness of the skip connection mechanism from different perspectives. In this work, we deep dive into the model's behaviors with skip connections which can be formulated as a learnable Markov chain. An efficient Markov chain is preferred as it always maps the input data to the target domain in a better way. However, while a model is explained as a Markov chain, it is not guaranteed to be optimized following an efficient Markov chain by existing SGD-based optimizers which are prone to get trapped in local optimal points. In order to towards a more efficient Markov chain, we propose a simple routine of penal connection to make any residual-like model become a learnable Markov chain. Aside from that, the penal connection can also be viewed as a particular model regularization and can be easily implemented with one line of code in the most popular deep learning frameworks~\footnote{Source code: \url{https://github.com/densechen/penal-connection}}. The encouraging experimental results in multi-modal translation and image recognition empirically confirm our conjecture of the learnable Markov chain view and demonstrate the superiority of the proposed penal connection., Comment: 12 pages, 4 figures

For years, the YOLO series has been the de facto industry-level standard for efficient object detection. The YOLO community has prospered overwhelmingly to enrich its use in a multitude of hardware platforms and abundant scenarios. In this technical report, we strive to push its limits to the next level, stepping forward with an unwavering mindset for industry application. Considering the diverse requirements for speed and accuracy in the real environment, we extensively examine the up-to-date object detection advancements either from industry or academia. Specifically, we heavily assimilate ideas from recent network design, training strategies, testing techniques, quantization, and optimization methods. On top of this, we integrate our thoughts and practice to build a suite of deployment-ready networks at various scales to accommodate diversified use cases. With the generous permission of YOLO authors, we name it YOLOv6. We also express our warm welcome to users and contributors for further enhancement. For a glimpse of performance, our YOLOv6-N hits 35.9% AP on the COCO dataset at a throughput of 1234 FPS on an NVIDIA Tesla T4 GPU. YOLOv6-S strikes 43.5% AP at 495 FPS, outperforming other mainstream detectors at the same scale~(YOLOv5-S, YOLOX-S, and PPYOLOE-S). Our quantized version of YOLOv6-S even brings a new state-of-the-art 43.3% AP at 869 FPS. Furthermore, YOLOv6-M/L also achieves better accuracy performance (i.e., 49.5%/52.3%) than other detectors with a similar inference speed. We carefully conducted experiments to validate the effectiveness of each component. Our code is made available at https://github.com/meituan/YOLOv6., Comment: technical report

Panoptic Narrative Grounding (PNG) is an emerging task whose goal is to segment visual objects of things and stuff categories described by dense narrative captions of a still image. The previous two-stage approach first extracts segmentation region proposals by an off-the-shelf panoptic segmentation model, then conducts coarse region-phrase matching to ground the candidate regions for each noun phrase. However, the two-stage pipeline usually suffers from the performance limitation of low-quality proposals in the first stage and the loss of spatial details caused by region feature pooling, as well as complicated strategies designed for things and stuff categories separately. To alleviate these drawbacks, we propose a one-stage end-to-end Pixel-Phrase Matching Network (PPMN), which directly matches each phrase to its corresponding pixels instead of region proposals and outputs panoptic segmentation by simple combination. Thus, our model can exploit sufficient and finer cross-modal semantic correspondence from the supervision of densely annotated pixel-phrase pairs rather than sparse region-phrase pairs. In addition, we also propose a Language-Compatible Pixel Aggregation (LCPA) module to further enhance the discriminative ability of phrase features through multi-round refinement, which selects the most compatible pixels for each phrase to adaptively aggregate the corresponding visual context. Extensive experiments show that our method achieves new state-of-the-art performance on the PNG benchmark with 4.0 absolute Average Recall gains., Comment: Accepted by ACM MM 2022

Existing approaches to image captioning usually generate the sentence word-by-word from left to right, with the constraint of conditioned on local context including the given image and history generated words. There have been many studies target to make use of global information during decoding, e.g., iterative refinement. However, it is still under-explored how to effectively and efficiently incorporate the future context. To respond to this issue, inspired by that Non-Autoregressive Image Captioning (NAIC) can leverage two-side relation with modified mask operation, we aim to graft this advance to the conventional Autoregressive Image Captioning (AIC) model while maintaining the inference efficiency without extra time cost. Specifically, AIC and NAIC models are first trained combined with shared visual encoders, forcing the visual encoder to contain sufficient and valid future context; then the AIC model is encouraged to capture the causal dynamics of cross-layer interchanging from NAIC model on its unconfident words, which follows a teacher-student paradigm and optimized with the distribution calibration training objective. Empirical evidences demonstrate that our proposed approach clearly surpass the state-of-the-art baselines in both automatic metrics and human evaluations on the MS COCO benchmark. The source code is available at: https://github.com/feizc/Future-Caption., Comment: ACM Multimedia 2022

Referring video object segmentation aims to predict foreground labels for objects referred by natural language expressions in videos. Previous methods either depend on 3D ConvNets or incorporate additional 2D ConvNets as encoders to extract mixed spatial-temporal features. However, these methods suffer from spatial misalignment or false distractors due to delayed and implicit spatial-temporal interaction occurring in the decoding phase. To tackle these limitations, we propose a Language-Bridged Duplex Transfer (LBDT) module which utilizes language as an intermediary bridge to accomplish explicit and adaptive spatial-temporal interaction earlier in the encoding phase. Concretely, cross-modal attention is performed among the temporal encoder, referring words and the spatial encoder to aggregate and transfer language-relevant motion and appearance information. In addition, we also propose a Bilateral Channel Activation (BCA) module in the decoding phase for further denoising and highlighting the spatial-temporal consistent features via channel-wise activation. Extensive experiments show our method achieves new state-of-the-art performances on four popular benchmarks with 6.8% and 6.9% absolute AP gains on A2D Sentences and J-HMDB Sentences respectively, while consuming around 7x less computational overhead., Comment: Accepted by CVPR 2022

Recently, lane detection has made great progress with the rapid development of deep neural networks and autonomous driving. However, there exist three mainly problems including characterizing lanes, modeling the structural relationship between scenes and lanes, and supporting more attributes (e.g., instance and type) of lanes. In this paper, we propose a novel structure guided framework to solve these problems simultaneously. In the framework, we first introduce a new lane representation to characterize each instance. Then a topdown vanishing point guided anchoring mechanism is proposed to produce intensive anchors, which efficiently capture various lanes. Next, multi-level structural constraints are used to improve the perception of lanes. In the process, pixel-level perception with binary segmentation is introduced to promote features around anchors and restore lane details from bottom up, a lane-level relation is put forward to model structures (i.e., parallel) around lanes, and an image-level attention is used to adaptively attend different regions of the image from the perspective of scenes. With the help of structural guidance, anchors are effectively classified and regressed to obtain precise locations and shapes. Extensive experiments on public benchmark datasets show that the proposed approach outperforms state-of-the-art methods with 117 FPS on a single GPU., Comment: Accepted by IJCAI 2021

BiSeNet has been proved to be a popular two-stream network for real-time segmentation. However, its principle of adding an extra path to encode spatial information is time-consuming, and the backbones borrowed from pretrained tasks, e.g., image classification, may be inefficient for image segmentation due to the deficiency of task-specific design. To handle these problems, we propose a novel and efficient structure named Short-Term Dense Concatenate network (STDC network) by removing structure redundancy. Specifically, we gradually reduce the dimension of feature maps and use the aggregation of them for image representation, which forms the basic module of STDC network. In the decoder, we propose a Detail Aggregation module by integrating the learning of spatial information into low-level layers in single-stream manner. Finally, the low-level features and deep features are fused to predict the final segmentation results. Extensive experiments on Cityscapes and CamVid dataset demonstrate the effectiveness of our method by achieving promising trade-off between segmentation accuracy and inference speed. On Cityscapes, we achieve 71.9% mIoU on the test set with a speed of 250.4 FPS on NVIDIA GTX 1080Ti, which is 45.2% faster than the latest methods, and achieve 76.8% mIoU with 97.0 FPS while inferring on higher resolution images.

Food recognition plays an important role in food choice and intake, which is essential to the health and well-being of humans. It is thus of importance to the computer vision community, and can further support many food-oriented vision and multimodal tasks. Unfortunately, we have witnessed remarkable advancements in generic visual recognition for released large-scale datasets, yet largely lags in the food domain. In this paper, we introduce Food2K, which is the largest food recognition dataset with 2,000 categories and over 1 million images.Compared with existing food recognition datasets, Food2K bypasses them in both categories and images by one order of magnitude, and thus establishes a new challenging benchmark to develop advanced models for food visual representation learning. Furthermore, we propose a deep progressive region enhancement network for food recognition, which mainly consists of two components, namely progressive local feature learning and region feature enhancement. The former adopts improved progressive training to learn diverse and complementary local features, while the latter utilizes self-attention to incorporate richer context with multiple scales into local features for further local feature enhancement. Extensive experiments on Food2K demonstrate the effectiveness of our proposed method. More importantly, we have verified better generalization ability of Food2K in various tasks, including food recognition, food image retrieval, cross-modal recipe retrieval, food detection and segmentation. Food2K can be further explored to benefit more food-relevant tasks including emerging and more complex ones (e.g., nutritional understanding of food), and the trained models on Food2K can be expected as backbones to improve the performance of more food-relevant tasks. We also hope Food2K can serve as a large scale fine-grained visual recognition benchmark., Comment: Accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence

Optimizing the approximation of Average Precision (AP) has been widely studied for image retrieval. Limited by the definition of AP, such methods consider both negative and positive instances ranking before each positive instance. However, we claim that only penalizing negative instances before positive ones is enough, because the loss only comes from these negative instances. To this end, we propose a novel loss, namely Penalizing Negative instances before Positive ones (PNP), which can directly minimize the number of negative instances before each positive one. In addition, AP-based methods adopt a fixed and sub-optimal gradient assignment strategy. Therefore, we systematically investigate different gradient assignment solutions via constructing derivative functions of the loss, resulting in PNP-I with increasing derivative functions and PNP-D with decreasing ones. PNP-I focuses more on the hard positive instances by assigning larger gradients to them and tries to make all relevant instances closer. In contrast, PNP-D pays less attention to such instances and slowly corrects them. For most real-world data, one class usually contains several local clusters. PNP-I blindly gathers these clusters while PNP-D keeps them as they were. Therefore, PNP-D is more superior. Experiments on three standard retrieval datasets show consistent results with the above analysis. Extensive evaluations demonstrate that PNP-D achieves the state-of-the-art performance. Code is available at https://github.com/interestingzhuo/PNPloss

Food recognition has received more and more attention in the multimedia community for its various real-world applications, such as diet management and self-service restaurants. A large-scale ontology of food images is urgently needed for developing advanced large-scale food recognition algorithms, as well as for providing the benchmark dataset for such algorithms. To encourage further progress in food recognition, we introduce the dataset ISIA Food- 500 with 500 categories from the list in the Wikipedia and 399,726 images, a more comprehensive food dataset that surpasses existing popular benchmark datasets by category coverage and data volume. Furthermore, we propose a stacked global-local attention network, which consists of two sub-networks for food recognition. One subnetwork first utilizes hybrid spatial-channel attention to extract more discriminative features, and then aggregates these multi-scale discriminative features from multiple layers into global-level representation (e.g., texture and shape information about food). The other one generates attentional regions (e.g., ingredient relevant regions) from different regions via cascaded spatial transformers, and further aggregates these multi-scale regional features from different layers into local-level representation. These two types of features are finally fused as comprehensive representation for food recognition. Extensive experiments on ISIA Food-500 and other two popular benchmark datasets demonstrate the effectiveness of our proposed method, and thus can be considered as one strong baseline. The dataset, code and models can be found at http://123.57.42.89/FoodComputing-Dataset/ISIA-Food500.html., Comment: Accepted by ACM Multimedia 2020

Facial landmark localization is a very crucial step in numerous face related applications, such as face recognition, facial pose estimation, face image synthesis, etc. However, previous competitions on facial landmark localization (i.e., the 300-W, 300-VW and Menpo challenges) aim to predict 68-point landmarks, which are incompetent to depict the structure of facial components. In order to overcome this problem, we construct a challenging dataset, named JD-landmark. Each image is manually annotated with 106-point landmarks. This dataset covers large variations on pose and expression, which brings a lot of difficulties to predict accurate landmarks. We hold a 106-point facial landmark localization competition1 on this dataset in conjunction with IEEE International Conference on Multimedia and Expo (ICME) 2019. The purpose of this competition is to discover effective and robust facial landmark localization approaches., Comment: This paper is accepted at ICME2019 Grand Challenge. The JD-landmark dataset has been released and can be downloaded from https://sites.google.com/view/hailin-shi