Pushkar Protik Goswami ¹, Dushyant Kumar Singh ¹

Affiliations
1 MNNIT Allahabad, Allahabad - 211004, Uttar Pradesh, IN

Abstract

Objectives: A number of techniques available in literature do not discuss the problem of incorrect object detection in the scenes having unstable or moving background. A novel hybrid technique is proposed in the paper to cover this problem. Methods/Statistical Analysis: The new approach proposed is the hybrid of two well known techniques for object detection. One is frame-differencing approach and other is skin colour modelling. This newer technique exploits the fact that the demerit of one technique works as merit for other and hence hybrid technique resolves the problem of moving background or turbulence in background. A real time video having turbulence in background is used for testing the efficiency of the approach. Findings: With accuracy of 0.97, the proposed approach outperforms the individual approaches i.e. frame-differencing and skin colour modelling. A very low value of False Positive Rate (FPR) for proposed approach compared to other approaches confirms the least incorrect detections. High value of True Positive Rate (TPR) conveys that fall out of correct object is least in the video by proposed approach. Results show that the proposed approach better applicable for background with turbulence. Application/Improvements: Automatic object detection and tracking in applications as surveillance are tractable with the proposed hybrid approach.

Keywords

Frame-Differencing, Object Detection, Object Tracking, Skin Colour, Thresholding.

Full Text

   

Dushyant Kumar Singh ¹, Dharmender Singh Kushwaha ¹

Affiliations
1 Department of Computer Science and Engineering, MNNIT Allahabad - 211004, Uttar Pradesh, IN

Abstract

Objectives: Numerous techniques have been proposed in past for skin colour modelling and some out of these are utilized in human detection problem. This paper proposes a novel technique which uses Indexed Look-up-table (ILUT) for skin colour modelling. Methods: In the proposed technique, skin colour cluster on 2-dimensional Cb-Cr plane in YCbCr colour space is modelled with an ILUT. ILUT contains the lower and upper bounds of Cb values corresponding to each value of Cr in the skin colour cluster. Outliers in the cluster are removed by applying median filter, because they contribute to the wrong classification of skin colour and hence wrong detection. Findings: The proposed technique delivers reasonably good performance on True Positive Rate (TPR), False Positive Rate (FPR) and Accuracy parameters. Classification accuracy of 89.89% for proposed technique is almost comparable to that of other techniques in literature. Comparative results are shown in Table 2 of the paper. Classification complexity is one parameter on which the proposed technique outperforms the rest of the other skin colour modelling techniques and it is of O(MxN). MxN is the image resolution. Skin colour modelling using ILUT is also efficient in terms of space requirements as compared to other non-parametric methods of skin colour modelling. Applications: Least classification complexity makes this technique most appropriate for real time systems/ applications for detecting presence of human being.

Keywords

Classification Accuracy, Classification Complexity, Human Detection, Indexed Look-up-table, Skin Colour Modelling.

Full Text