U2-ONet: A Two-Level Nested Octave U-Structure Network with a Multi-Scale Attention Mechanism for Moving Object Segmentation

Most scenes in practical applications are dynamic scenes containing moving objects, so accurately segmenting moving objects is crucial for many computer vision applications. In order to efficiently segment all the moving objects in the scene, regardless of whether the object has a predefined semantic label, we propose a two-level nested octave U-structure network with a multi-scale attention mechanism, called U2-ONet. U2-ONet takes two RGB frames, the optical flow between these frames, and the instance segmentation of the frames as inputs. Each stage of U2-ONet is filled with the newly designed octave residual U-block (ORSU block) to enhance the ability to obtain more contextual information at different scales while reducing the spatial redundancy of the feature maps. In order to efficiently train the multi-scale deep network, we introduce a hierarchical training supervision strategy that calculates the loss at each level while adding knowledge-matching loss to keep the optimization consistent. The experimental results show that the proposed U2-ONet method can achieve a state-of-the-art performance in several general moving object segmentation datasets.

Download Full-text

Accurate moving object segmentation in unconstraint videos based on robust seed pixels selection

International Journal of Advanced Robotic Systems ◽

10.1177/1729881420947273 ◽

2020 ◽

Vol 17 (4) ◽

pp. 172988142094727

Author(s):

Wenlong Zhang ◽

Xiaoliang Sun ◽

Qifeng Yu

Keyword(s):

Optical Flow ◽

Moving Objects ◽

Object Segmentation ◽

Moving Object ◽

Data Sets ◽

Seed Selection ◽

Current Frame ◽

Moving Object Segmentation ◽

Random Walker ◽

Previous Frame

Due to the clutter background motion, accurate moving object segmentation in unconstrained videos remains a significant open problem, especially for the slow-moving object. This article proposes an accurate moving object segmentation method based on robust seed selection. The seed pixels of the object and background are selected robustly by using the optical flow cues. Firstly, this article detects the moving object’s rough contour according to the local difference in the weighted orientation cues of the optical flow. Then, the detected rough contour is used to guide the object and the background seed pixel selection. The object seed pixels in the previous frame are propagated to the current frame according to the optical flow to improve the robustness of the seed selection. Finally, we adopt the random walker algorithm to segment the moving object accurately according to the selected seed pixels. Experiments on publicly available data sets indicate that the proposed method shows excellent performance in segmenting moving objects accurately in unconstraint videos.

Download Full-text

AN ADAPTIVE BACKGROUND UPDATION AND GRADIENT BASED SHADOW REMOVAL METHOD

International Journal of Image Processing and Vision Science ◽

10.47893/ijipvs.2013.1057 ◽

2013 ◽

pp. 288-293

Author(s):

SUMIT KUMAR SINGH ◽

MAGAN SINGH

Keyword(s):

Background Subtraction ◽

Moving Objects ◽

Object Segmentation ◽

Moving Object ◽

Shadow Removal ◽

Subtraction Method ◽

Moving Object Segmentation ◽

Shadow Elimination ◽

Illumination Changes ◽

Gradient Based

Moving object segmentation has its own niche as an important topic in computer vision. It has avidly being pursued by researchers. Background subtraction method is generally used for segmenting moving objects. This method may also classify shadows as part of detected moving objects. Therefore, shadow detection and removal is an important step employed after moving object segmentation. However, these methods are adversely affected by changing environmental conditions. They are vulnerable to sudden illumination changes, and shadowing effects. Therefore, in this work we propose a faster, efficient and adaptive background subtraction method, which periodically updates the background frame and gives better results, and a shadow elimination method which removes shadows from the segmented objects with good discriminative power. Keywords- Moving object segmentation,

Download Full-text

Particle Swarm Clustering-Based Moving Object Segmentation in the H.264 Compressed Domain

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.433-440.4841 ◽

2012 ◽

Vol 433-440 ◽

pp. 4841-4844

Author(s):

Pei Wang ◽

Jing Wang

Keyword(s):

Clustering Algorithm ◽

Moving Objects ◽

Object Segmentation ◽

Median Filter ◽

Particle Swarm ◽

Moving Object ◽

Compressed Domain ◽

Motion Vectors ◽

Moving Object Segmentation ◽

Vector Median

An approach of the moving object segmentation is proposed in this paper. Firstly the motion fields are extracted from the compressed stream, where the noise and the unreal motion blocks are removed by vector median filter. Then the motion vectors are accumulated by motion estimation, in order to get denser and prominent motion vectors. Finally the moving objects are segmented adaptively by particle swarm clustering algorithm. It is demonstrated by the experimental results that the moving objects in the compressed domain can be segmented effectively.

Download Full-text

Research on Information Technology with Moving Object Segmentation Based on Memory Matrix and Kalman Filter

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.859.482 ◽

2013 ◽

Vol 859 ◽

pp. 482-485

Author(s):

Can Can Zhou ◽

Xiao Run Li

Keyword(s):

Information Technology ◽

Kalman Filter ◽

Moving Objects ◽

Object Segmentation ◽

Region Growing ◽

Moving Object ◽

Video Sequences ◽

Multiple Channels ◽

Segmentation Method ◽

Moving Object Segmentation

In this paper,information technology is introduced briefly and in order to avoid inaccurate segmentation of moving objects caused by object holes and ghost, an automatic moving object segmentation method which belongs to information technology based on memory matrix and Kalman filter theory is proposed. Memory matrix is used in multiple channels to extract the initial background, and ghost is eliminated after updating background according to the theory of Kalman filter. Moving objects are extracted using adaptive threshold, and object segmentation is achieved by improved region growing method on base of block processing. The experimental results indicate that the proposed algorithm can accurately segment moving object from video sequences, and has very good robustness against illumination variance and moving noise.

Download Full-text

Symmetry Encoder-Decoder Network with Attention Mechanism for Fast Video Object Segmentation

Symmetry ◽

10.3390/sym11081006 ◽

2019 ◽

Vol 11 (8) ◽

pp. 1006 ◽

Cited By ~ 2

Author(s):

Mingyue Guo ◽

Dejun Zhang ◽

Jun Sun ◽

Yiqi Wu

Keyword(s):

Object Segmentation ◽

Target Object ◽

General Purpose ◽

Attention Mechanism ◽

Video Object Segmentation ◽

Video Object ◽

Feature Maps ◽

Multi Scale ◽

Object Mask ◽

Fine Tune

Semi-supervised video object segmentation (VOS) has obtained significant progress in recent years. The general purpose of VOS methods is to segment objects in video sequences provided with a single annotation in the first frame. However, many of the recent successful methods heavily fine-tune the object mask in the first frame, which decreases their efficiency. In this work, to address this issue, we propose a symmetry encoder-decoder network with the attention mechanism for video object segmentation (SAVOS) requiring only one forward pass to segment the target object in a video. Specifically, the encoder generates a low-resolution mask with smoothed boundaries, while the decoder further refines the details of the segmentation mask and integrates lower level features progressively. Besides, to obtain accurate segmentation results, we sequentially apply the attention module on multi-scale feature maps for refinement. We conduct several experiments on three challenging datasets (i.e., DAVIS 2016, DAVIS 2017, and SegTrack v2) to show that SAVOS achieves competitive performance against the state-of-the-art.

Download Full-text