Detection and Analysis of Sow Targets Based on Image Vision

In large-scale sow production, real-time detection and recognition of sows is a key step towards the application of precision livestock farming techniques. In the pig house, the overlap of railings, floors, and sows usually challenge the accuracy of sow target detection. In this paper, a non-contact machine vision method was used for sow targets perception in complex scenarios, and the number position of sows in the pen could be detected. Two multi-target sow detection and recognition models based on the deep learning algorithms of Mask-RCNN and UNet-Attention were developed, and the model parameters were tuned. A field experiment was carried out. The data-set obtained from the experiment was used for algorithm training and validation. It was found that the Mask-RCNN model showed a higher recognition rate than that of the UNet-Attention model, with a final recognition rate of 96.8% and complete object detection outlines. In the process of image segmentation, the area distribution of sows in the pens was analyzed. The position of the sow’s head in the pen and the pixel area value of the sow segmentation were analyzed. The feeding, drinking, and lying behaviors of the sow have been identified on the basis of image recognition. The results showed that the average daily lying time, standing time, feeding and drinking time of sows were 12.67 h(MSE 1.08), 11.33 h(MSE 1.08), 3.25 h(MSE 0.27) and 0.391 h(MSE 0.10), respectively. The proposed method in this paper could solve the problem of target perception of sows in complex scenes and would be a powerful tool for the recognition of sows.

AOGAN: A generative adversarial network for screen space ambient occlusion

Ambient occlusion (AO) is a widely-used real-time rendering technique which estimates light intensity on visible scene surfaces. Recently, a number of learning-based AO approaches have been proposed, which bring a new angle to solving screen space shading via a unified learning framework with competitive quality and speed. However, most such methods have high error for complex scenes or tend to ignore details. We propose an end-to-end generative adversarial network for the production of realistic AO, and explore the importance of perceptual loss in the generative model to AO accuracy. An attention mechanism is also described to improve the accuracy of details, whose effectiveness is demonstrated on a wide variety of scenes.

NeRF

We present a method that achieves state-of-the-art results for synthesizing novel views of complex scenes by optimizing an underlying continuous volumetric scene function using a sparse set of input views. Our algorithm represents a scene using a fully connected (nonconvolutional) deep network, whose input is a single continuous 5D coordinate (spatial location ( x , y , z ) and viewing direction ( θ, ϕ )) and whose output is the volume density and view-dependent emitted radiance at that spatial location. We synthesize views by querying 5D coordinates along camera rays and use classic volume rendering techniques to project the output colors and densities into an image. Because volume rendering is naturally differentiable, the only input required to optimize our representation is a set of images with known camera poses. We describe how to effectively optimize neural radiance fields to render photorealistic novel views of scenes with complicated geometry and appearance, and demonstrate results that outperform prior work on neural rendering and view synthesis.

Exploration of the Relationships between Hazard Perception and Eye Movement for Young Drivers

Aiming at young drivers’ hazard perception (HP) and eye movement, a cross-sectional study was conducted in the city of Xi’an, China. 46 participants were recruited, and 35 traffic scenes were used to test drivers’ hazard perception and eye movement. The difference analysis and correlation analysis were carried out for the acquired data. The results suggest that some indices of hazard perception and eye movement are significantly correlated. A higher saccade speed is in the direction of higher hazardous scenes. Higher complex scenes result in smaller saccade angle. The number of hazards unidentified is negatively influenced by complexity degree and hazardous degree of traffic scenes, and similar associations are found between hazard identification time, complexity degree, and hazardous degree. The hazard identification time and the number of hazards slowly identified are positively affected by the number of fixations and the number of saccades. Meanwhile, differences in the hazardous degree evaluation, hazard identification time, number of hazards unidentified, number of fixations, and number of saccades are found in different types of traffic scenes. The results help us to improve the design of road and vehicle devices, as well as the assessment and enhancement of young drivers’ hazard perception skills.

Efficient attention-based deep encoder and decoder for automatic crack segmentation

Recently, crack segmentation studies have been investigated using deep convolutional neural networks. However, significant deficiencies remain in the preparation of ground truth data, consideration of complex scenes, development of an object-specific network for crack segmentation, and use of an evaluation method, among other issues. In this paper, a novel semantic transformer representation network (STRNet) is developed for crack segmentation at the pixel level in complex scenes in a real-time manner. STRNet is composed of a squeeze and excitation attention-based encoder, a multi head attention-based decoder, coarse upsampling, a focal-Tversky loss function, and a learnable swish activation function to design the network concisely by keeping its fast-processing speed. A method for evaluating the level of complexity of image scenes was also proposed. The proposed network is trained with 1203 images with further extensive synthesis-based augmentation, and it is investigated with 545 testing images (1280 × 720, 1024 × 512); it achieves 91.7%, 92.7%, 92.2%, and 92.6% in terms of precision, recall, F1 score, and mIoU (mean intersection over union), respectively. Its performance is compared with those of recently developed advanced networks (Attention U-net, CrackSegNet, Deeplab V3+, FPHBN, and Unet++), with STRNet showing the best performance in the evaluation metrics-it achieves the fastest processing at 49.2 frames per second.

SRSDD-v1.0: A High-Resolution SAR Rotation Ship Detection Dataset

Deep learning has been widely used in the field of SAR ship detection. However, current SAR ship detection still faces many challenges, such as complex scenes, multiple scales, and small targets. In order to promote the solution to the above problems, this article releases a high-resolution SAR ship detection dataset which can be used for rotating frame target detection. The dataset contains six categories of ships. In total, 30 panoramic SAR tiles of the Chinese Gaofen-3 of port areas with a 1-m resolution were cropped to slices, each with 1024 × 1024 pixels. In addition, most of the images in the dataset contain nearshore areas with complex background interference. Eight state-of-the-art rotated detectors and a CFAR-based method were used to evaluate the dataset. Experimental results revealed that the complex background will have a great impact on the performance of detectors.

Global Illumination on a Mobile Phone: Scalable Real-time Global Illumination using Sparse Radiance Probes

<p>Real-time global illumination that scales from low to high-end hardware is important for interactive applications so they can reach wider audiences. To do this, the real-time lighting algorithm used needs to have varying performance characteristics. Sparse Radiance Probes (SRP) is a recent real-time global illumination algorithm that runs in under 5 ms per frame on a high-end Nvidia Titan X GPU. Its low per-frame timings suggest it could scale to low-end devices, but no prior work provides complete implementation details and evaluates its performance across devices with varying performance characteristics to prove this. Therefore, this thesis aims to fill this gap and determine if SRP is scalable across low to high-end devices. SRP is implemented with adjustable scaling parameters, and its performance is compared across three test devices. A low-end iPhone 7, a mid-range AMD Radeon 560 graphics card, and a high-end AMD RX Vega 56 graphics card. The implementation in this thesis ran above 60 FPS for simple scenes on the iPhone 7, and with a reasonable reduction in quality, it ran just above 30 FPS on more complex scenes like Crytek Sponza. These results show that SRP can scale to low-end devices. While the implementation in this thesis runs in real time, there are implementation optimisations that would make SRP run even faster across all the test devices without reducing quality.</p>

Global Illumination on a Mobile Phone: Scalable Real-time Global Illumination using Sparse Radiance Probes

<p>Real-time global illumination that scales from low to high-end hardware is important for interactive applications so they can reach wider audiences. To do this, the real-time lighting algorithm used needs to have varying performance characteristics. Sparse Radiance Probes (SRP) is a recent real-time global illumination algorithm that runs in under 5 ms per frame on a high-end Nvidia Titan X GPU. Its low per-frame timings suggest it could scale to low-end devices, but no prior work provides complete implementation details and evaluates its performance across devices with varying performance characteristics to prove this. Therefore, this thesis aims to fill this gap and determine if SRP is scalable across low to high-end devices. SRP is implemented with adjustable scaling parameters, and its performance is compared across three test devices. A low-end iPhone 7, a mid-range AMD Radeon 560 graphics card, and a high-end AMD RX Vega 56 graphics card. The implementation in this thesis ran above 60 FPS for simple scenes on the iPhone 7, and with a reasonable reduction in quality, it ran just above 30 FPS on more complex scenes like Crytek Sponza. These results show that SRP can scale to low-end devices. While the implementation in this thesis runs in real time, there are implementation optimisations that would make SRP run even faster across all the test devices without reducing quality.</p>

Research and Implementation of Real-time Render Optimization Algorithm Based on GPU

In the continuous optimization of computer graphics, game engine and virtual reality have become the focus of research and innovation in the current technology field. Rendering is a core technology to show a variety of graphic effects, which has been attached great importance to and applied by the whole society. Nowadays, although the rendering of large-scale and complex scenes has a wide range of applications, the actual optimization requirements are very high. Therefore, in the future technology application and research and development, based on higher and higher technical requirements, it is necessary to improve the efficiency of actual rendering while ensuring or improving the rendering quality. Therefore, this paper studies how to use the algorithm to improve the efficiency of actual rendering, so as to provide a new basis for future computer graphics research.