In order to explore the application of the image recognition model based on multi-stage convolutional neural network (MS-CNN) in the deep learning neural network in the intelligent recognition of commodity images and the recognition performance of the method, in the study, the features of color, shape, and texture of commodity images are first analyzed, and the basic structure of deep convolutional neural network (CNN) model is analyzed. Then, 50,000 pictures containing different commodities are constructed to verify the recognition effect of the model. Finally, the MS-CNN model is taken as the research object for improvement to explore the influence of label errors (p = 0.03, 0.05, 0.07, 0.09, 0.12) with different parameter settings and different probabilities (size of convolutional kernel, Dropout rate) on the recognition accuracy of MS-CNN model, at the same time, a CIR system platform based on MS-CNN model is built, and the recognition performance of salt and pepper noise images with different SNR (0, 0.03, 0.05, 0.07, 0.1) was compared, then the performance of the algorithm in the actual image recognition test was compared. The results show that the recognition accuracy is the highest (97.8%) when the convolution kernel size in the MS-CNN model is 2*2 and 3*3, and the average recognition accuracy is the highest (97.8%) when the dropout rate is 0.1; when the error probability of picture label is 12%, the recognition accuracy of the model constructed in this study is above 96%. Finally, the commodity image database constructed in this study is used to identify and verify the model. The recognition accuracy of the algorithm in this study is significantly higher than that of the Minitch stochastic gradient descent algorithm under different SNR conditions, and the recognition accuracy is the highest when SNR = 0 (99.3%). The test results show that the model proposed in this study has good recognition effect in the identification of commodity images in scenes of local occlusion, different perspectives, different backgrounds, and different light intensity, and the recognition accuracy is 97.1%. To sum up, the CIR platform based on MS-CNN model constructed in this study has high recognition accuracy and robustness, which can lay a foundation for the realization of subsequent intelligent commodity recognition technology.
The immune-excluded tumors (IETs) show limited response to current immunotherapy due to intrinsic and adaptive immune resistance. In this study, it is identified that inhibition of transforming growth factor-β (TGF-β) receptor 1 can relieve tumor fibrosis, thus facilitating the recruitment of tumor-infiltrating T lymphocytes. Subsequently, a nanovesicle is constructed for tumor-specific co-delivery of a TGF-β inhibitor (LY2157299, LY) and the photosensitizer pyropheophorbide a (PPa). The LY-loaded nanovesicles suppress tumor fibrosis to promote intratumoral infiltration of T lymphocytes. Furthermore, PPa chelated with gadolinium ion is capable of fluorescence, photoacoustic and magnetic resonance triple-modal imaging-guided photodynamic therapy, to induce immunogenic death of tumor cells and elicit antitumor immunity in preclinical cancer models in female mice. These nanovesicles are further armored with a lipophilic prodrug of the bromodomain-containing protein 4 inhibitor (i.e., JQ1) to abolish programmed death ligand 1 expression of tumor cells and overcome adaptive immune resistance. This study may pave the way for nanomedicine-based immunotherapy of the IETs.
The traditional key frame extraction algorithm can't effectively segment the surveillance video, and it can't focus on the moving objects in the surveillance video data. In this paper, a key frame extraction algorithm based on golden section is proposed. The method first detects and tracks the moving target in the surveillance video, and calculates the entropy value of the foreground image. Secondly, the golden section in mathematical calculation is introduced to divide the sub-segment of the surveillance video, and the standard deviation of the foreground image entropy is used to measure the intra-frame similarity of the video sub-segments. If the intra-frame similarity is low, the frame at the golden section point is selected as the video key frame;if the difference within the video segment is large, golden section is continued until there is no significant difference in the video frames in all video sub-segments. in all the video sub-segments. Through experimental tests on a variety of surveillance video data and comparison with traditional algorithms, the experimental results show that the proposed algorithm effectively compresses the original surveillance video, and can extract the moving objects in the surveillance video more completely.
Achieving selective and durable inhibition of programmed death ligand 1 (PD-L1) in tumors for T cell activation remains a major challenge in immune checkpoint blockade therapy. We herein presented a set of clickable inhibitors for spatially confined PD-L1 degradation and radioimmunotherapy of cancer. Using metabolic glycan engineering click bioorthogonal chemistry, PD-L1 expressed on tumor cell membranes was labeled with highly active azide groups. This enables covalently binding of the clickable inhibitor with PD-L1 and subsequent PD-L1 degradation. A pH-activatable nanoparticle responding to extracellular acidic pH of tumor was subsequently used to deliver the clickable PD-L1 inhibitor into extracellular tumor microenvironment for depleting PD-L1 on the surface of tumor cell and macrophage membranes in vivo. We further demonstrated that a combination of the clickable PD-L1 inhibitor with radiotherapy (RT) eradicated the established tumor by inhibiting RT–up-regulated PD-L1 in the tumor tissue. Therefore, selective PD-L1 blockade in tumors via the clickable PD-L1 inhibitor offers a versatile approach to promote cancer immunotherapy.
The purpose of this research is to explore the optimization and fusion application of multimodal neuroimaging technology and analyze the evaluation method of human brain fatigue based on multimodal neuroimaging technology. Based on electroencephalogram (EEG) and fMRI (functional magnetic resonance imaging), the four‐dimensional consistency of local neural activities (FOCA) and local multimodal serial analysis (LMSA) are first introduced to fuse EEG and fMRI organically. Second, the eigenspace maximal information canonical correlation analysis (emiCCA) is introduced to construct the multimodal neuroimaging data fusion system. Finally, how the brain function network is constructed is introduced. Based on the binary and the weighted brain function networks, the relationship between the human brain fatigue and the brain function network is evaluated by calculating the fractal dimension. Results demonstrate that FOCA performs well in temporal and spatial consistency indexes, and the mean level and standard deviation in the case of temporal and spatial consistency are approximately 0.45. The effect of LMSA indexes is significantly better than generalized linear models (GLMs). Under different signal‐to‐noise ratios (SNRs), the regression coefficient based on LMSA is much larger than the GLM estimate; the corresponding significance level is p < 0.05; and the maximum value of the regression coefficient appears near 0.2. In the data fusion system, the time‐space matching has good results under the time accuracy based on EEG and the space accuracy based on fMRI, with the time accuracy above 88% and the space accuracy above 89%. The fractal dimension analysis based on the brain function network reveals that the weighted brain function network is more sensitive to mental fatigue. The state of human brain fatigue will make the brain function network more complicated. The fractal dimension with more network edges is around 2.2, while the fractal dimension with fewer network edges is around 1.6. The proposed data analysis and fusion system have great application potential and propose a new idea for analyzing human brain fatigue and brain aging.
Forgery detection is essential to verify the integrity and authenticity of images. Existing block-based detection techniques detect forgery in the same image, most of which use similar frameworks while differ in the feature extraction schemes. These methods have high accuracy in detecting the forged regions, but the computational load is heavy when facing exhaustive search problems. This paper describes a forgery detection method based on local binary pattern residue classes and color regions. An image is divided into overlapped blocks. Local binary pattern residue classes are computed for each block. The plane formed by a dimensional and b dimensional from Lab color space is divided into 16 regions. Similar blocks are searched in the overlapped blocks with the same local binary pattern residue class and color region, then they are grouped into several suspicious regions. Finally, we analyze the multi-region relation of these suspicious regions and their areas to locate the tampered regions. The small hole is filled through the morphologic operation. The results of experiments demonstrated that our method has good performance in that it improved detection accuracy and reduced execution time under various challenging conditions. As the proposed method reduces the search range for similar blocks, it has a higher speed than exhaustive search and has comparable detection results at the same time.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)