Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{191903,
        author = {para rajesh and P Sushanth Chakravarthy and Neha Sheetal Kodakandla and B Siddhu and T Manoj Kumar},
        title = {DeepVision: An Interpretable Multi-Task Framework for Comprehensive Bone Fracture Diagnosis Using ResNet, DenseNet, YOLOv8, and Grad-CAM++},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {8},
        pages = {8381-8388},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=191903},
        abstract = {Bones break X- rays In many things everyday medicine. Secure it fast helps patients Secure better faster. Most AI tools Just do it now one thing- Value spotting or tidying up- and often just for the sake of it one body part. It does them hard Implement as directed by physicians broad solutions.
To fix these problems, this study brings out DeepVision- One step by step deep learning setup Built to place, sort, determine and pin broken bones I seven body areas: shoulders, upper arm, elbow, lower arm, Wrist, hand, fingers. It packs a punch four Tuned pieces: ResNet34 with a CBAM block This highlights the main features for the break. DenseNet121 Also CBAM to inform different break kinds In addition; YOLOv8n To fasten the breach without delay; And Grad- CAM++ To manifest exactly where the damage is.
Model power is extended using pre- trained networks, Additional synthetic images, With cleaning of input scans before processing. Performance checks depend on common measures- Favor correctness rate, Average detection score( map), and what the final shape is prefer the marked spots to do real ones- preserve results trustworthy. Administer tests known collections Esteem MURA and the Kaggle Bone Fracture set So we can understand how well it works new cases.
DeepVision Mix sharp detection power with clean visuals- Spit out label, box outline about box breaks, Plus shade- coded hot zone Shows Locations of damage. That way, Doctors can locate out what's causing it the system Something The flag instead of just zooming justice one area Favor older tools do, this setup Handle multiple regions At the same time minimize errors and perform well different bone types.
It is designed to be elementary to perform heavy computing needs, to create it handy to real clinics.
Sure, there are still some hiccups- Prefer to make fun of tiny cracks or uneven training data- But overall, This can increase how quickly and reliably it breaks down bones Procure Stuck in scans, for which no- nonsense helper offered skeletal imaging.},
        keywords = {Bone fracture detection, Multi- level deep learning, X- ray image analysis, Resnett 34, DenseNet121, CBAM( Convolutional Block Attention Module), YOLOv8, Grade Chem++, Fracture classification, object recognition, Medical image localization, Multidimensional muscle imaging, computer- aided diagnosis( CAD), transfer learning, Medical image processing},
        month = {February},
        }