Deep infrared pedestrian classification based on automatic image matting

Liang, Yihui; Huang, Han; Cai, Zixing; Zhang, Hao; Tan, Kay Chen

doi:10.1016/j.asoc.2019.01.024

Cited by 21 publications

(14 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The brighter the position indicates its higher activation level and the better the feature extraction. As shown in Figure 1, showing the heat map of network layer 55 and network layer 23 in yolov4 [6]. By observing Figure 1, few pedestrians are activated in Figure 1b, and relatively more pixel points are activated in the background.…”

Section: Heat Map Based Npe Construction Methodsmentioning

confidence: 96%

“…With the widespread use of computer vision, deep learning networks based on multimodal data (visible images (RGB images), depth images, infrared images) are applied to various fields, including object detection [1][2][3][4][5] and classification [6,7], image segmentation [8][9][10], target tracking [11][12][13], etc. While most of the mainstream algorithms are designed based on RGB images and show better performance when extracting key features from RGB images.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Pruning Method for Deep Convolutional Network Based on Heat Map Generation Metrics

Zhang

Wang

Chen

et al. 2022

Sensors

View full text Add to dashboard Cite

With the development of deep learning, researchers design deep network structures in order to extract rich high-level semantic information. Nowadays, most popular algorithms are designed based on the complexity of visible image features. However, compared with visible image features, infrared image features are more homogeneous, and the application of deep networks is prone to extracting redundant features. Therefore, it is important to prune the network layers where redundant features are extracted. Therefore, this paper proposes a pruning method for deep convolutional network based on heat map generation metrics. The ‘network layer performance evaluation metrics’ are obtained from the number of pixel activations in the heat map. The network layer with the lowest ‘network layer performance evaluation metrics’ is pruned. To address the problem that the simultaneous deletion of multiple structures may result in incorrect pruning, the Alternating training and self-pruning strategy is proposed. Using a cyclic process of pruning each model once and retraining the pruned model to reduce the incorrect pruning of network layers. The experimental results show that proposed method in this paper improved the performance of CSPDarknet, Darknet and Resnet.

show abstract

Section: Heat Map Based Npe Construction Methodsmentioning

confidence: 96%

Section: Introductionmentioning

confidence: 99%

A Pruning Method for Deep Convolutional Network Based on Heat Map Generation Metrics

Zhang

Wang

Chen

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…The aim of natural image matting is to determine the alpha values of pixels in unknown regions, while those in known foreground and known background regions are denoted as 1 and 0, respectively. Natural image matting has been applied not only to mid-level vision tasks such as image fusion [1], automatic foreground extraction [2], [3], semantic segmentation [4], medical image processing [5] and target tracking [6], but also to high-level vision tasks, such as pedestrian classification [7], object recognition [2], virtual reality [8], and street view augmented reality [9]. As the range of image matting applications broadened, the image matting approach is required to provide satisfactory alpha mattes within a given amount of computing resources.…”

Section: Introductionmentioning

confidence: 99%

“…The pixel-pair-optimization-based approach is one of the competitive image matting approaches that have distinct advantages in parallelization [10] and in handling a mislabeled trimap [7] or spatially disconnected foreground [11]- [13]. In this approach, natural image matting problem is modeled as a pixel pair optimization (PPO) problem that can be written as: min g p (x p ) s.t.…”

Section: Introductionmentioning

confidence: 99%

Pyramid Matting: A Resource-Adaptive Multi-Scale Pixel Pair Optimization Framework for Image Matting

2020

Self Cite

View full text Add to dashboard Cite

Image matting is an important problem in computer vision with significant theoretical interest and diverse practical applications, including image/video editing, target tracking, and object recognition. Pixel-pair-optimization-based image matting approaches have been shown very successful in estimating the opacity of the foreground by searching for the best pair of foreground and background pixels for each unknown pixel. However, extant approaches encounter difficulties in adapting to the changes of available computing resources, which limits the application of image matting. This drawback has motivated the present study, as a part of which a multi-scale evolutionary pixel pair optimization framework named pyramid matting framework (PMF) was developed. In this framework, the large-scale pixel pair optimization problem is transformed to multiple pixel pair optimization problems of different scales using image pyramid. The resulting problems are solved level by level, starting from the problem at the small scale. Pixel pair heuristic information obtained from solving low-scale problems are iteratively propagated to the spatially-related pixel pairs in the larger-scale problem. PMF can adapt to changes in available computing resources due to its capability of transforming a small-scale problem solution to the large-scale problem solution through the heuristic information propagation. Experimental results show that the PMF-based image matting approach not only provides high-quality alpha mattes with sufficient computing resources, but also works well when computing resources are scarce.

show abstract

“…Many efforts have been made to develop shape features that are not sensitive to brightness, orientation, or scale. However, the generalization capacity of expert-driven learning approaches is often very limited, and some of them are designed to solve specific problems [ 15 ]. For instance, Bassem et al distinguished pedestrians from other objects by utilizing a SURF-based feature based on the assumption that a pedestrian’s head appears as a light region [ 16 ].…”

Section: Introductionmentioning

confidence: 99%

Feature Channel Expansion and Background Suppression as the Enhancement for Infrared Pedestrian Detection

Wang

et al. 2020

Sensors

View full text Add to dashboard Cite

Pedestrian detection is an important task in many intelligent systems, particularly driver assistance systems. Recent studies on pedestrian detection in infrared (IR) imagery have employed data-driven approaches. However, two problems in deep learning-based detection are the implicit performance and time-consuming training. In this paper, a novel channel expansion technique based on feature fusion is proposed to enhance the IR imagery and accelerate the training process. Besides, a novel background suppression method is proposed to stimulate the attention principle of human vision and shrink the region of detection. A precise fusion algorithm is designed to combine the information from different visual saliency maps in order to reduce the effect of truncation and miss detection. Four different experiments are performed from various perspectives in order to gauge the efficiency of our approach. The experimental results show that the Mean Average Precisions (mAPs) of four different datasets have been increased by 5.22% on average. The results prove that background suppression and suitable feature expansion will accelerate the training process and enhance the performance of IR image-based deep learning models.

show abstract

Deep infrared pedestrian classification based on automatic image matting

Cited by 21 publications

References 21 publications

A Pruning Method for Deep Convolutional Network Based on Heat Map Generation Metrics

A Pruning Method for Deep Convolutional Network Based on Heat Map Generation Metrics

Pyramid Matting: A Resource-Adaptive Multi-Scale Pixel Pair Optimization Framework for Image Matting

Feature Channel Expansion and Background Suppression as the Enhancement for Infrared Pedestrian Detection

Contact Info

Product

Resources

About