Search

Denoising Diffusion Probabilistic Model (DDPM) is able to make flexible conditional image generation from prior noise to real data, by introducing an independent noise-aware classifier to provide conditional gradient guidance at each time step of denoising process. However, due to the ability of classifier to easily discriminate an incompletely generated image only with high-level structure, the gradient, which is a kind of class information guidance, tends to vanish early, leading to the collapse from conditional generation process into the unconditional process. To address this problem, we propose two simple but effective approaches from two perspectives. For sampling procedure, we introduce the entropy of predicted distribution as the measure of guidance vanishing level and propose an entropy-aware scaling method to adaptively recover the conditional semantic guidance. For training stage, we propose the entropy-aware optimization objectives to alleviate the overconfident prediction for noisy data.On ImageNet1000 256x256, with our proposed sampling scheme and trained classifier, the pretrained conditional and unconditional DDPM model can achieve 10.89% (4.59 to 4.09) and 43.5% (12 to 6.78) FID improvement respectively. The code is available at https://github.com/ZGCTroy/ED-DPM., 24 pages, 8 figures

With the advent of depth sensors, the use of RGB and depth(D) information for salient object detection (SOD) has been explored extensively in recent years. However, the depth quality from the different scenes usually varies, leading to fusing and achieving complementary between RGB and low-quality depth is still a challenging problem. In this paper, we first design a Double Dilated Merge Module (DDMM) to extract comprehensive and beneficial high-level cross-modality features and explore further global context information at multi-scales to obtain a coarse saliency map. Then, we propose a Cross-Modality Enhance Module (CMEM) to enhance cross-modality features compatibility and fuse them with the previous predicted coarse saliency map to generate a more accurate saliency map. Furthermore, we introduce a Distribution-Region Combination Loss (DRcom Loss) to optimize our proposed Double Cross-Modality Progressively Guided Network (DCPGNet) in a coarse-to-fine manner. DCPGNet achieves satisfactory performance on five public benchmarks compared with recent state-of-the-art algorithms.

Satellite orbit error is one of main sources in Global Positioning System (GPS) receiver positioning error. Using broadcast ephemeris in navigation message, the receiver can calculate spatial position of GPS satellites, then combine pseudorange measurements and achieve positioning solution. In this work, taking broadcast ephemeris in the 2036th GPS week as an example, orbital position of the satellite with the pseudo-random-noise (PRN) code 4 is calculated. Furthermore, using GPS precision ephemeris as a reference, orbital error of the broadcast ephemeris for the PRN 4 satellite is investigated. The results show that the maximum difference of the satellite in three directions is 36 meters, 57 meters and 39 meters, respectively. The three-dimension distance deviating from precision ephemeris is at most 61 meters. The maximum deviation occurs only during certain time periods in the whole GPS week.

Blends of polycarbonate (PC) and acrylonitrile-ethylene-propylene-diene-styrene (AES) were reactive compatibilized by styrene-maleic anhydride copolymers (SMA). The changes in phase morphology and interfacial characteristics of the blends as a function of maleic anhydride content of SMA and the concentration of compatibilizer have been systematic studied. The occurrence of reaction between the terminal hydroxyl groups of PC and the maleic anhydride (MA) of compatibilizer was confirmed by fourier transform infrared (FTIR) spectroscopy. A glass transition temperature (Tg) with an intermediate value between Tg(AES) and Tg(PC) was found on differential scanning calorimeter (DSC) curves of PC/AES blends compatibilized with SMA contains high levels of MA. Furthermore, at lower compatibilizer content, increase of the compatibilizer level in blends result in decreasing gap between two Tgs corresponding to the constituent polymers. Small angle X-ray scattering (SAXS) test results indicated that compatibilizer concentration for the minimum of blend interface layer's thickness was exactly the same as it was when compatibilized PC/AES blend exhibited optimal compatibility in DSC test. The observed morphological changes were consistent well with the DSC and SAXS test results. A new mechanism of interfacial structural development was proposed to explain unusual phenomena of SMA compatibilized PC/AES blends. © 2015 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2015, 132, 42103.