Ta-SLOA: Taylor snow leopard optimization with XCovNet-based object classification using vehicle image

Detecting objects is a core challenge in computer vision and plays a vital role in applications like self-driving cars, traffic surveillance, and smart transportation networks. Accurate vehicle object detection is essential for various applications such as autonomous driving, traffic surveillance, and intelligent transportation systems. This research proposes an optimization-based deep learning (DL) model for object classification from vehicle images. Initially, the input image is denoised by employing the adaptive weighted median filter (AWMF). Next, the contrast limited adaptive histogram equalization (CLAHE) is used for enhancing the image, which improves its quality. Thereafter, the fast fuzzy c-means (FFCM) clustering model is used for segmentation of objects and then augmentation of images is done. The segmented image is then subjected to object detection using You Only Look Once v9 Squeeze M-SegNet (YOLO v9-S Net), which is obtained by combining Squeeze M-SegNet (SM-SegNet) with YOLO v9 models. Finally, object classification is effectuated by Xception Convolutional Neural Network (XCovNet), and the Taylor snow leopard optimization algorithm (Ta-SLOA) is employed for training the XCovNet model. Ta-SLOA is a novel algorithmic technique devised by merging Taylor series and snow leopard optimization algorithm (SLOA). The Ta-SLOA_XCovNet outperforms traditional techniques, achieving impressive precision, F1-score, and recall values of 93.181%, 93.724%, and 92.970%. Experimental results demonstrate that the proposed Ta-SLOA-optimized XCovNet framework significantly improves vehicle object detection accuracy and efficiency, thereby offering a robust solution for real-time applications in autonomous driving and intelligent transportation systems.