Image & Vision

Investigates Image data fusion techniques that combine image and track data from multiple sensors to achieve improved accuracies and more specific inferences than could be achieved by the use of a single sensor alone. Our aim is to explore the state-of-the-art image processing algorithms for achieving effective data fusion as in:

Preprocessing and Detection

작성자 관리자 날짜 2021-04-08 14:51:10 조회수 139


1. Preprocessing       

The key function of preprocessing  is to improve the image in ways that increase the chances for success of the other processes. Preprocessing typically deals with techniques for enhancing contrast, removing noise, and isolating regions.


2. Face Detection      

The face detection procedure consists of two stages. First, the skin color regions are segmented based on the chrominance component of the input image. Then the regions with elliptical shape are selected as face hypothesis. The face region is confirmed by searching for the facial features in their interior. The skin color alone is not sufficient so we combine skin color and shape to find the face candidate.

3. Car Detection (Segmentation)      

The first step in image analysis generally is to segment the image. Segmentation subdivides an image into its constinuent parts or objects. The level to which this subdivision in carried depends on the problem being solved. That is, segmentation should stop when the objects of interest in an application have been isolated.

4.  Hand Detection   

Hand detection and tracking is very important technique for hand gesture recognition. Because we need to recognize continuous motions of hand. Detecion finds the area of hand, and  tracking is keeping that area without missing. Following demo is the tracking the area of hand in case of several hand candidates.

5. Wide Dynamic Range (WDR)     

Wide Dynamic Range (WDR) is the technique for back-light compensation (BLC). For example, security camera needs the clear images to recognize objects in any environment, such as back-light. General image sensor, CCD or CMOS, is not better than human’s eyes in dynamic range. Therefore, for general image sensor, it is very useful for image-fusion method used several images with different exposure time. Luminance Fusion use Gradient-domain and chroma fusion use merge by regions.

댓글 (0)

등록된 댓글이 없습니다.
작성 권한이 없습니다.