Chi-Chia Sun

Fast Motion Object Detection Algorithm Using Complementary Depth Image on an RGB-D Camera

Stereo vision has become a popular topic in recent years, especially in-depth images from stereo vision. Depth information can be extracted either from a dual camera or RGB-D camera. In image processing, the realization of object detection is only based on the color information or depth images separately; however, both have advantages and disadvantages. Therefore, many researchers have combined them together to achieve better results. A new fast motion object-detection algorithm is presented based on the complementary depth images and color information, which is able to detect motion objects without background noise. The experiment results show that the proposed fast object detection algorithm can achieve 84.4% of the segmentation accuracy rate on average with a 45 FPS computation speed on an embedded platform.

Low-cost MCU solution for measuring HRV using SpO2 sensor

In this paper, a low-cost Heart Beat Variability (HRV) analyzer base on the principle of Photoplethysmography (PPG) is presented on an ultra-low power MCU platform. It is a non-invasive method of measuring the variation in blood volume in tissues using a light source and a detector. Base on the blood variation, the proposed HRV device will be able to calculate the heart beat rate, level of oxygenation and HRV data. In emergency, the specific information will be send back to the ICU center through the 3G-WCDMA interface, where caregivers can further analyze it and determine treatment methods.

Object detection using adaptive block-based background model

This paper propose an adaptable block-based background modeling and real time image object detection algorithm. In training step, we present adaptable block-based background model that uses major color number to determine the block size. This background model can reduce the memory consumption, efficiently. In detection step, we use one pixel to compare with background model. Then, it can reduce processing time. The experiment results show that we can save 33.9% memory space. Finally, we can achieve 27.25 frames per second for the benchmark video with image size 768×576.

Efficient object motion detection based on RGB-D image

Low-cost RGB-D camera, such as Kinect have widely used in image processing. Although Kinect can provide depth and color information simultaneously, but the depth information is very noisy. In this paper, we proposed a novel motion detection scheme based on edge-and-depth to enhance the object motion detection in RGB-D camera. Our proposed algorithm can fast detect motion regions without background model building. It is efficient to be used in many applications, e.g. image recognition, tracking.

Low-power microcontroller solution for measuring HBR using single reflection SpO2 Sensor

In this paper, a low-power Heart Beat Rate (HBR) analyzer based on the principle of single reflective Photoplethysmography (PPG) is presented on a MCU platform. It is a non-invasive method of measuring the variation in blood volume in tissues using a light source and a light detector. The PPG signal is amplified and converted to digital format by AFE (Analog Front End) device, MCU takes the signal via SPI communication and performs HBR finding algorithm, afterward sending results to the mobile phone through the Bluetooth 4.0 wireless communication.