Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Filters

20242

Settings

Has files: NoSubject: search.filters.subject.Homography

Search Results

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Enhanced Video Mosaic Generation: Efficient ORB Feature Extraction and Hamming Distance Matching
    (Institute of Electrical and Electronics Engineers Inc., 2024) Shridhar, H.; Harakannanavar, S.S.; Chetan, R.; Kanabur, V.; Jayalaxmi, H.; Prashanth, C.R.
    Video mosaicing is a technique for creating expansive panoramic views by seamlessly combining multiple video frames. This paper introduces an innovative algorithm for video mosaic generation, emphasizing the alignment and blending of non-overlapping frames. The proposed algorithm utilizes the ORB feature extraction technique to effectively address challenges related to camera motion and content variations across frames, ensuring a smooth and continuous mosaic output. Key aspects of the algorithm include optimizing the number of feature matches for frame stitching, which critically impacts the final mosaic's quality. Experimental results demonstrate the algorithm's ability to produce visually coherent and high-quality video mosaics. The model has been tested on real-time video frames, showing improved performance of 95% of accuracy with lower RMSE and contributing to the advancement of video mosaicing techniques for applications in surveillance and other fields requiring comprehensive scene views. © 2024 IEEE.
  • No Thumbnail Available
    Item
    Optimized Image Mosaicing Using ORB with Histogram Equalization for Real-Time Applications
    (Institute of Electrical and Electronics Engineers Inc., 2024) Shridhar, H.; Harakannanavar, S.S.; Chetan, R.; Kanabur, V.; Jayalaxmi, H.; Prashanth, R.C.
    Image mosaicing combines overlapping images of the same scene into a larger, seamless image. This work aims to develop a model that efficiently merges images while evaluating its performance in terms of runtime and the number of key features used. Histogram Equalization (HE) is applied to handle intensity variability, while key features are extracted using SIFT, ORB, and BRISK descriptors. K-Nearest Neighbor (KNN) matches the features, and homography is estimated using the RANSAC algorithm to align the images. A smoothing filter is applied to create the final panorama. Experimental results show that ORB with HE is the most efficient technique, using 500 key features and achieving a runtime of 0.0836 seconds. The average runtime of the entire image mosaicing process, from feature extraction to panorama generation, was significantly lower for ORB than for SIFT (0.45 seconds) and BRISK (0.30 seconds). The minimal runtime and reduced computational cost of ORB make it ideal for applications requiring real-time image stitching, such as video processing or live panoramic imaging. The proposed model addresses limitations in feature matching and offers improved performance in image mosaicing. © 2024 IEEE.