Search

The purpose of this study was to assess the feasibility of testing binocular visual acuity using the automated acuity card procedure (AACP)-a new automated method that uses an eye-tracking system. We included participants aged 5 to 36 months old. Binocular grating acuity was tested using the AACP and Teller Acuity Cards (TACs) with a uniform testing distance (55 cm) in random order. Electronic stimuli in the AACP were similar in size and form to TACII and roughly equivalent to the printed TACII stimuli. The AACP tracked the participant's gaze from the recorded image sequences and automatically determined the grating acuity. Differentiation, correlation, and consistency were compared between the AACP and TACs. Ultimately, 77.11% (155/201) of participants completed both tests. Fewer participants failed the TAC test than the AACP (15 [7.46%] vs. 31 [15.42%]). The average duration of the AACP (median interquartile range [IQR] = 60 [IQR = 41] seconds) was significantly shorter than that of the TAC (median [IQR], 185 [IQR = 66] seconds, p0.001). AACP and TAC visual acuities were robustly correlated (r = 0.83, p0.001). Bland-Altman plots revealed a mean difference between AACP and TAC visual acuities of 0.10 cycles per degree (cpd; 95% limits of agreement = 7.70 cpd). Both the AACP and TACs indicated improved visual acuity with age progression (both, p0.001), with no significant differences between the tests. Electronic stimuli were presented using the AACP yielded clinically useful data on grating acuity in infants and toddlers. AACP performance was comparable to that of TACs, the current clinical gold standard for assessing infant vision regarding testability, reliability, and accuracy.

Conventional deep learning based methods for object detection require a large amount of bounding box annotations for training, which is expensive to obtain such high quality annotated data. Few-shot object detection, which learns to adapt to novel classes with only a few annotated examples, is very challenging since the fine-grained feature of novel object can be easily overlooked with only a few data available. In this work, aiming to fully exploit features of annotated novel object and capture fine-grained features of query object, we propose Dense Relation Distillation with Context-aware Aggregation (DCNet) to tackle the few-shot detection problem. Built on the meta-learning based framework, Dense Relation Distillation module targets at fully exploiting support features, where support features and query feature are densely matched, covering all spatial locations in a feed-forward fashion. The abundant usage of the guidance information endows model the capability to handle common challenges such as appearance changes and occlusions. Moreover, to better capture scale-aware features, Context-aware Aggregation module adaptively harnesses features from different scales for a more comprehensive feature representation. Extensive experiments illustrate that our proposed approach achieves state-of-the-art results on PASCAL VOC and MS COCO datasets. Code will be made available at https://github.com/hzhupku/DCNet., Accepted by CVPR2021

Fresh fruit bruising caused by mechanical impact is the most problematic effect of mechanical postharvest processing of tomatoes. The conventional postharvest process of tomato fruit (Solanum lycopersicum L.) handling was surveyed from harvest to shipping preparation at a farm in Toechon in the Republic of Korea. Acceleration sensors located on the harvest boxes and fruit skin showed that the most severe impact forces (>10 g) occurred while fruit was moved in and out of the harvest box. Next, multivariate analysis was used to evaluate nondestructively the susceptibility of tomatoes to bruising by developing five estimation models incorporating quality factors and the presence of bruising resulting from peak contact forces. Linear regression models, artificial neural network (ANN) regression models, and a logistic regression model were built; the primary dependent variables were the rates of weight loss and firmness loss, and impact-induced bruising. The impact force was controlled by using a pendulum, which produced four levels of impact force on tomato fruit. Increasing the relative humidity (RH) and firmness, and decreasing the temperature and degree of weight loss decreased tomato fruit bruise damage. In addition, the proposed drop–impact measurement technique, with multivariate analysis, can be used to evaluate tomato quality nondestructively.

Sparse representation-based methods have been successfully applied to visual tracking. However, complex and inefficient optimization limits their deployment in practical tracking scenarios. In this paper, we propose a temporal consistency dictionary learning tracking algorithm to enable efficient dictionary learning and tracking executive. First, we present an objective function which introduces the fixed dictionary and variance dictionary to reconstruct the object’s appearance. In particular, the proposed method takes the temporal consistency into account by adding a regularization term into the objective function to constrain the difference of object appearance at adjacent frames. Then the optimization problem is solved in an iteration way. Moreover, the proposed method can encode the object’s local structural information, and the local patches from the same candidate altogether for a global appearance representation. Second, we develop an effective observation likelihood function based on the proposed model. It takes the influence of patches with large reconstruction errors into consideration, thereby, alleviating the drifting of the object. Finally, we present an appearance updating strategy to adapt to the object’s appearance variations by the online dictionary learning. Experimental evaluations on the TB50 and TB100 datasets show that the proposed tracking method outperforms sparse representation related visual tracking as well as other state-of-the-art tracking methods.

This study was conducted to investigate the weight loss, firmness, external color and vitamin C (VC) content of tomatoes (Lycopersicon esculentum) using non-destructive method to measure identical tomato samples during 15 days storage at low temperature and high humidity. Tomatoes were harvested at the light red stage, sorted, box packed and then stored in thermo-hygrostat (10±1°C, 90±10%RH). The quality changes in weight loss, firmness and external color were measured every 3 day interval. Weight loss was increased by 1.13±0.15%, but it may not be considered to affect quality. Surface color of fruit was changed, especially in lightness and hue angle value. The color values were analyzed by analysis of variance (ANOVA), and the results were significant (p

Discoloration and dehydration are crucial problems associated with quality of bell pepper (Capsicum annuum L.). Color and weight loss of the bell pepper were measured at different vibration levels, storage environments and storage periods. Vibration tests were carried out at different ‘g’ levels from normal to the worst-case for 180 sec on the initial day. The effect of two different storage environments (10°C, 90 % RH and 18°C, 40 % RH) for four time periods (1, 4, 7 and 10 days) were observed. At 18°C, 40 % RH storage environment, more weight loss and duller color were observed. In the 10°C, 90 % RH storage environment, the group of bell pepper sample having 12 g level vibration had significantly much more weight loss compared to other groups. In this storage environment, most of bell pepper samples maintained vivid color. Storage life and fruit quality of bell pepper were affected by storage temperatures, humidity and duration. The results showed that high g level vibration, high temperature, and low humidity environment during harvesting and storage process of the bell pepper should be avoided.

Near-infrared spectroscopy and multivariate analysis techniques were employed to nondestructively evaluate the rancidity of perilla seed oil by developing prediction models for the acid and peroxide values. The acid, peroxide value, and transmittance spectra of perilla seed oil stored in two different environments for 96 and 144 h were obtained and used to develop prediction models for different storage conditions and time periods. Preprocessing methods were applied to the transmittance spectra of perilla seed oil, and multivariate analysis techniques, such as principal component regression (PCR), partial least squares regression (PLSR), and artificial neural network (ANN) modeling, were employed to develop the models. Titration analysis shows that the free fatty acids in an oil oxidation process were more affected by relative humidity than temperature, whereas peroxides in an oil oxidation process were more significantly affected by temperature than relative humidity for the two different environments in this study. Also, the prediction results of ANN models for both acid and peroxide values were the highest among the developed models. These results suggest that the proposed near-infrared spectroscopy technique with multivariate analysis can be used for the nondestructive evaluation of the rancidity of perilla seed oil, especially the acid and peroxide values.

Effective video-based detection methods are of great importance to intelligent transportation systems (ITS), and here we propose a method to localize and label objects. The method is able to detect pedestrians and bicycle riders in a complex scene. Our method is inspired by the common fate principle, which is a mechanism of visual perception in human beings, and which states tokens moving or functioning in a similar manner tend to be perceived as one unit. Our method embeds the principle in an Implicit Shape Model (ISM). In our method, keypoint-based object parts are firstly detected and then grouped by their motion patterns. Based on the grouping results, when the object parts vote for object centers and labels, each vote belonging to the same object part is assigned a weight according to its consistency with the votes of other object parts in the same motion group. Afterwards, the peaks, which correspond to detection hypotheses on the Hough image formed by summing up all weighted votes, become easier to find. Thus our method performs better in both position and label estimations. Experiments show the effectiveness of our method in terms of detection accuracy.

Tracking generic human motion is highly challenging due to its high-dimensional state space and the various motion types involved. In order to deal with these challenges, a fusion formulation which integrates low- and high-dimensional tracking approaches into one framework is proposed. The low-dimensional approach successfully overcomes the high-dimensional problem of tracking the motions with available training data by learning motion models, but it only works with specific motion types. On the other hand, although the high-dimensional approach may recover the motions without learned models by sampling directly in the pose space, it lacks robustness and efficiency. Within the framework, the two parallel approaches, low- and high-dimensional, are fused via a probabilistic approach at each time step. This probabilistic fusion approach ensures that the overall performance of the system is improved by concentrating on the respective advantages of the two approaches and resolving their weak points. The experimental results, after qualitative and quantitative comparisons, demonstrate the effectiveness of the proposed approach in tracking generic human motion.

Laser-based people tracking systems have been developed for mobile robotic, and intelligent surveillance areas. Existing systems rely on laser point clustering method to extract object locations. However, for dense crowd tracking, laser points of different objects are often interlaced and undistinguishable due to measurement noise and they can not provide reliable features. It causes current systems quite fragile and unreliable. This paper presents a novel and robust laser-based dense crowd tracking method. Firstly, we introduce a stable feature extraction method based on accumulated distribution of successive laser frames. With this method, the noise that generates split and merged measurements is smoothed away and the pattern of rhythmic swing legs is utilized to extract each leg of persons. And then, a region coherency property is introduced to construct an efficient measurement likelihood model. The final tracker is based on the combination of independent Kalman filter and Rao-Blackwellized Monte Carlo data association filter (RBMC-DAF). In real experiments, we obtain raw data from multiple registered laser scanners, which measure two legs for each people on the height of 16cm from horizontal ground. Evaluation with real data shows that the proposed method is robust and effective. It achieves a significant improvement compared with existing laser-based trackers. In addition, the proposed method is much faster than previous works, and can overcome tracking errors resulted from mixed data of two closely situated persons.