OPEN SOURCE NETWORK SOLUTION COMMUNITY
X-RAY MACHINE PROTOPYE
We coordinated with SHT joint stock company and The Da Nang IT Infrastructure Development Center to implement a research plan for low cost digital x-ray machines. With the support from Star Dynamic Corp., New Jersey, USA, with the free license of a patent, we have built a comparable digital X-ray machine. With expensive digital cameras and only about a tenth of the price of imported products
GENIHOG OBJECT DETECTOR
We are developing GENIHOS OBJECT DETECTOR as an IP core, designed and developed for embedded traffic surveillance applications. The inside algorithm is based on a HOG descriptor and a SVM classifier. The Histogram of Oriented Gradients (HOG) is an algorithm commonly used in computer vision and image processing for the purpose of object recognition. Support Vector Machine (SVM) is a non probabilistic binary linear model that analyze data for classification.
The SVM model is off-line trained based on two sets of image sample which is relatively motorcycle and non-motorcycle. They are extracted from real traffic videos with wide range scenarios. The result of SVM model parameter is used for IP core design. The core can support multiple scales to detect motorcycle moving in a wide range of distance.
GENIBS MOTION DETECTOR
The GeniBS Motion Detector is an IP core, designed and developed for moving object detection application. The core is the combination of Background Subtraction (BS) and Shadow Removal. The BS algorithm, based on Pixel Based Adaptive Segmented (PBAS) allowed detecting moving objects. The Shadow Removal algorithm mostly get rid of moving cast shadows which is classified as unnecessary, so that the motion detection can be improved more accurately.
The background model in PBAS is the buffer of N recently frames. The foreground/background classiﬁcation is made on comparing the current frame with these N frames. The update is applied randomly on samples with a speciﬁed probability.