Chaobo Min

Unsupervised evaluation method using Markov random field for moving object segmentation in infrared videos

An unsupervised method is proposed for performance evaluation of the moving object segmentation using Markov random field (MRF) in infrared videos. This method focuses on the edge features and takes spatio-temporal information into account. The authors consider an MRF model for each edge point of a segmentation mask in spatial and temporal directions. This problem is then formulated using maximum a posteriori estimation principle to form a criterion of evaluation. Subjective evaluation is applied to measure the performance of the evaluation methods. The results show that the proposed method is superior to other unsupervised measures.

PERFORMANCE EVALUATION OF INFRARED MOVING FOREGROUND SEGMENTATION WITHOUT PRIOR KNOWLEDGE

Moving foreground segmentation has been widely used for many infrared video applications. In this work, a metric is presented to evaluate the performance of infrared moving foreground segmentation without prior knowledge. Difference image is given by temporal difference and is quantized into two classes which can differentiate between changed pixel and fixed pixel. In the experiments, we find that, in the segmented image of infrared moving foreground, the difference between the spatial distributions of classes of segmented foreground and background can accurately reflect the difference between segmented foreground and background. So, we can claim if this difference is great, moving foreground segmentation is better. Experiments are performed on variety situation of moving foreground segmentation, showing that the proposed metric is effective.

Algorithm selective method of infrared/visual image fusion system based on air humidity

To find out the best infrared and visible fusion system of fusion algorithm which has excellent target detection characteristics in different environment, we proposed a new fusion algorithm selective rule. We also defined new concepts: fusion algorithm coefficient and the equivalent transmissivity of system. Using local-target contrast, local-target articulation to calculate fusion algorithm coefficient, we can estimate the target detection performance of fusion system when it working in different air humidity environment. Also, we make use of infrared and visible fusion system designed by ourselves to verify this method. Besides fusion algorithm coefficient, we also use subjective evaluation to evaluate the target detection performance of fusion algorithm. At last, the best algorithm or the method which is most consistent with human visual in different conditions were found. Through this work, we can provide the basis for the algorithm of choice in the fusion system.

New method for unsupervised segmentation of moving objects in infrared videos

Abstract. A new method for unsupervised segmentation of moving objects in infrared videos is presented. This method consists of two steps: difference image quantization and spatial segmentation. In the first step, the changed pixels in the difference image are quantized to several classes by using Bayes decision. It can be used to cluster the changed pixels belonging to the same moving object together. The pixels of the difference image are replaced by their corresponding class labels, thus forming a class-map of the difference image. In the second step, each class in the class-map is considered as a subset of the possible seeds of moving objects. A self-adaptive region growing method is then used to image segmentation on the basis of these different subsets. One of the focuses of this work is on spatial segmentation, where a criterion is proposed for evaluation of moving object segmentation without ground truth in infrared videos. This criterion is used to evaluate the performance of the segmentation masks grown from different subsets of the possible seeds. The best segmented image is determined to be the final segmentation result. Experiments show the advantage and robustness of the proposed algorithm on real infrared videos.

Intelligent Eye-Information Fusion Monitioring System

We have designed a new type of real-time image fusion system which named Intelligent Eye-Information Fusion monitoring system. The system achieves the functions of image fusion, target detection, image stitching. Meanwhile it can applicable to the field of security monitoring, and water search and rescue. The system mainly consists of two parts. One part is the detection system; another part is display system and control circuit. These two parts are connected by the transmission system. The first part is mainly used for multi-spectral image acquisition and image fusion, and the second part is mainly used to control the first part and display the image. We use two algorithms which are applied in the system. The one is used in grayscale fusion, and the other is used in color fusion. At last, four kinds of features are explained, including image fusion, image stitching, moving target detection, and switching function of visual field.