Nowadays, the field of long-range thermal imageprocessing is increasingly popular. Although this technology has several advantages in terms of security and monitoring, there are not enough autonomous algorithms for p...
详细信息
On end-to-end driving, human driving demonstrations are used to train perception-based driving models by imitation learning. This process is supervised on vehicle signals (e.g., steering angle, acceleration) but does ...
详细信息
ISBN:
(纸本)9781665491907
On end-to-end driving, human driving demonstrations are used to train perception-based driving models by imitation learning. This process is supervised on vehicle signals (e.g., steering angle, acceleration) but does not require extra costly supervision (human labeling of sensor data). As a representative of such vision-based end-to-end driving models, CILRS is commonly used as a baseline to compare with new driving models. So far, some latest models achieve better performance than CILRS by using expensive sensor suites and/or by using large amounts of human-labeled data for training. Given the difference in performance, one may think that it is not worth pursuing vision-based pure end-to-end driving. However, we argue that this approach still has great value and potential considering cost and maintenance. In this paper, we present CIL++, which improves on CILRS by both processing higher-resolution images using a human-inspired HFOV as an inductive bias and incorporating a proper attention mechanism. CIL++ achieves competitive performance compared to models which are more costly to develop. We propose to replace CILRS with CIL++ as a strong vision-based pure end-to-end driving baseline supervised by only vehicle signals and trained by conditional imitation learning.
A novel hyperspectral image classification algorithm is proposed and demonstrated on benchmark hyperspectral images. We also introduce a hyperspectral sky imaging dataset that we are collecting for detecting the amoun...
详细信息
Spatial-wise dynamic convolution has become a promising approach to improving the inference efficiency of deep networks. By allocating more computation to the most informative pixels, such an adaptive inference paradi...
详细信息
ISBN:
(纸本)9781713871088
Spatial-wise dynamic convolution has become a promising approach to improving the inference efficiency of deep networks. By allocating more computation to the most informative pixels, such an adaptive inference paradigm reduces the spatial redundancy in image features and saves a considerable amount of unnecessary computation. However, the theoretical efficiency achieved by previous methods can hardly translate into a realistic speedup, especially on the multi-core processors (e.g. GPUs). The key challenge is that the existing literature has only focused on designing algorithms with minimal computation, ignoring the fact that the practical latency can also be influenced by scheduling strategies and hardware properties. To bridge the gap between theoretical computation and practical efficiency, we propose a latency-aware spatial-wise dynamic network (LASNet), which performs coarse-grained spatially adaptive inference under the guidance of a novel latency prediction model. The latency prediction model can efficiently estimate the inference latency of dynamic networks by simultaneously considering algorithms, scheduling strategies, and hardware properties. We use the latency predictor to guide both the algorithm design and the scheduling optimization on various hardware platforms. Experiments on image classification, object detection and instance segmentation demonstrate that the proposed framework significantly improves the practical inference efficiency of deep networks. For example, the average latency of a ResNet-101 on the imageNet validation set could be reduced by 36% and 46% on a server GPU (Nvidia Tesla-V100) and an edge device (Nvidia Jetson TX2 GPU) respectively without sacrificing the accuracy. Code is available at https://***/LeapLabTHU/LASNet.
Agriculture sector is an important pillar of the global economy. The cotton crop is considered one of the prominent agricultural resources. It is widely cultivated in India, China, Pakistan, USA, Brazil, and other cou...
详细信息
ISBN:
(纸本)9781665462198
Agriculture sector is an important pillar of the global economy. The cotton crop is considered one of the prominent agricultural resources. It is widely cultivated in India, China, Pakistan, USA, Brazil, and other countries of the world. The worldwide cotton crop production is severely affected by numerous diseases such as cotton leaf curl virus (CLCV/CLCuV), bacterial blight, and ball rot. imageprocessing techniques together with machine learning algorithms are successfully employed in numerous fields and have also used for crop disease detection. In this study, we present a deep learning-based method for classifying diseases of the cotton crop, including bacterial blight and cotton leaf curl virus (CLCV). The dataset of cotton leaves showing disease symptoms is collected from various locations in Sindh, Pakistan. We employ the Inception v4 architecture as a convolutional neural network to identify diseased plant leaves in particular bacterial blight and CLCV. The accuracy of the designed model is 98.26% which shows prominent improvement compared to the existing models and systems.
The proceedings contain 33 papers. The topics discussed include: MBAPIS: multi-level behavior analysis guided program interval selection for microarchitecture studies;automatic code generation for high-performance gra...
ISBN:
(纸本)9798350342543
The proceedings contain 33 papers. The topics discussed include: MBAPIS: multi-level behavior analysis guided program interval selection for microarchitecture studies;automatic code generation for high-performance graph algorithms;SimplePIM: a software framework for productive and efficient processing-in-memory;Drishyam: an image is worth a data prefetcher;architecture-aware currying;PreFlush: lightweight hardware prediction mechanism for cache line flush and writeback;retargeting applications for heterogeneous systems with the tribble source-to-source framework;dynamic allocation of processor cores to graph applications on commodity servers;parallelizing maximal clique enumeration on GPUs;and HugeGPT: storing guest page tables on host huge pages to accelerate address translation.
The proceedings contain 71 papers. The topics discussed include: a new sampling strategy to improve the performance of mobile robot path planning algorithms;machine learning based methods for Arabic duplicate question...
ISBN:
(纸本)9781665495585
The proceedings contain 71 papers. The topics discussed include: a new sampling strategy to improve the performance of mobile robot path planning algorithms;machine learning based methods for Arabic duplicate question detection;robust traffic signs classification using deep convolutional neural network;image-based visual servoing techniques for robot control;SASHA: a shift-add segmented hybrid approximated multiplier for imageprocessing;graph based method for Arabic text summarization;face information forensics analysis based on facial aging: a survey;advanced financial data processing and labeling methods for machine learning;and recognition system of human activities based on time-frequency features of accelerometer data.
This research study aims to develop a pneumonia detection system using vision transformers. Pneumonia is a very serious respiratory illness that may result in severe health issues, and early detection is essential for...
详细信息
This article presents a new image segmentation algorithm based on a Split & Merge approach. By nature, the execution time of Split & Merge algorithms is data-dependent, as their halting conditions are tied to ...
详细信息
ISBN:
(数字)9798350349399
ISBN:
(纸本)9798350349405
This article presents a new image segmentation algorithm based on a Split & Merge approach. By nature, the execution time of Split & Merge algorithms is data-dependent, as their halting conditions are tied to the homogeneity of each region. While previous algorithms made the Split step less sensitive to input data, the execution time of the more complex Merge step remains highly sensitive to image content. This paper tackles the sensitivity and performance problems from a system and architecture perspective. Memory reallocations due to array fusions are eliminated with the introduction of a TTA (Three Table Array) structure in the Merge step. As iterating over entries in this structure causes a loss of memory locality, we propose two new mechanisms that implement a software cache to mitigate this. An experimental study on an embedded system (Nvidia Jetson Xavier NX) has shown our Merge algorithm to be 10.6 times faster than the state-of-the-art Split & Merge algorithm for $960 \times 720$ images. Moreover, the execution time of our algorithm is also more resistant to image characteristics.
Direction-of-arrival estimation is a significant problem in many telecommunication systems which their architecture is composed with antenna arrays. In literature, several algorithms were proposed as a challenge, for ...
详细信息
暂无评论