A Non-invasive blood Glucose Monitoring (NGM) method is desired for diabetes patients. Unfortunately, there is no practical NGM technique available for considering clinical accuracy requirements. One of the major caus...
详细信息
ISBN:
(纸本)9781728118680
A Non-invasive blood Glucose Monitoring (NGM) method is desired for diabetes patients. Unfortunately, there is no practical NGM technique available for considering clinical accuracy requirements. One of the major causes is raised from the interferences of Environmental Background (EB) and Physiological Background (PB) noises. In this study, we propose a novel NGM method, called NIV-NGM. Based on the near-infrared video of a user's finger, NIV-NGM automatically separates the blood vessel and the background (other tissues) by an image segmentation algorithm and estimates the glucose level using features extracted from them. Compared with other NGM methods, NIV-NGM can differentiate the glucose-related information from the background, and also compensates for the measurement results due to the background noises. Clinical experimental results demonstrate that NIV-NGM is robust and immune from EB and PB noise influence, while can achieve the expected monitoring accuracy (Clarke (A) = 0.96).
Low visibility weather has seriously affected daily traffic management and safety of life and property. Aiming at the problem that highway in low visibility environment is prone to traffic accidents, a highway visibil...
详细信息
ISBN:
(纸本)9781728123264
Low visibility weather has seriously affected daily traffic management and safety of life and property. Aiming at the problem that highway in low visibility environment is prone to traffic accidents, a highway visibility detection method based on video surveillance is proposed, and the surveillance video data of each section are analyzed. Based on the traditional dark channel prior, the highway visibility detection method converted the surveillance video frame into gray image, and extracted the air transmittance and atmospheric light value contained in the surveillance image by combining the adaptive guided filtering algorithm and quadtree image segmentation algorithm. Then based on the visibility calculation model, the visibility distance of the road was estimated according to the actual distance and offset angle between the monitoring equipment and the detection object. The experimental results show that the optimized image processing method can solve the problems of low image clarity, color distortion, and poor scene adaptability. The detection of visibility error rate meets the accuracy requirements of highway visibility detection in China. Finally, a road real-time visibility detection and warning system is applied based on the corresponding early warning strategies of each visibility level.
This paper presents a novel vision system design for the Rubik's cube *** vision system is implemented on an ARM9 *** s3c2440 development board was cut to meet hardware system,including the core-board,the bottom b...
详细信息
ISBN:
(纸本)9781457702686
This paper presents a novel vision system design for the Rubik's cube *** vision system is implemented on an ARM9 *** s3c2440 development board was cut to meet hardware system,including the core-board,the bottom board,touch screen,USB camera and other *** principle of the vision system can be described as follows:Six image of the Rubik's cube are captured by the USB camera. Each block on the cube is segmented and determined via a proposed image segmentation algorithm and *** then, an artificial intelligence search reduction algorithm is introduced to restore the Rubik's cube.
This paper describes an extension to the graph cut interactive image segmentation algorithm based on a novel approach to addressing the well known small cut problem. The approach uses a generative contrast model to we...
详细信息
ISBN:
(纸本)9781479923427
This paper describes an extension to the graph cut interactive image segmentation algorithm based on a novel approach to addressing the well known small cut problem. The approach uses a generative contrast model to weight interaction potentials. The model attempts to capture the expected changes in color between adjacent pixels in the unlabeled area of the image using the adjacent pixels in the user interactions as training data. We compare our approach to the standard graph cuts algorithm and show that the contrast model allows a user to achieve a more accurate segmentation with fewer interactions. We additionally introduce a variant of the approach based on superpixels that further enhances performance but reduces computational complexity to ensure instant feedback for optimal user experience.
暂无评论