An Optimal Approach to Detect the Human Heads using H-MTF in Crowded Scenes
Due to the increase in the number of people at crowded places leads to some disaster events, there is a necessity to detect the human heads and estimate the crowd density. The counting of the human heads is quite an immense topic in computer vision and digital image processing. This paper focuses on sample frames that are to be extracted from the crowd video UCF_HDDC and S_HOCK datasets. Our proposed Hybridization-Multiple Target Features (H-MTF) method, detects head objects using three prominent features: texture, color, and shape (T, C, and S). With the help of H-MTF, the optimal value can be estimated to detect the exact spot of the head in a crowded place. By applying two evaluation metrics: (i) Average Precision metric (AvP) and (ii) Average Recall metric (AvR), H-MTF has been compared with the existing methods using 2 different datasets. The results are shown in terms of AvP and AvR and our H-MTF method outcomes best from the existing methods.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.