Scene Analysis based on Feature Extraction and Deep Learning
Scene classification relies on images and is essential in many computer vision applications. The scientific interest in scene classification from remotely collected images is increasing, and many datasets and algorithms are being developed. The introduction of convolutional neural networks (CNN) and
2025-06-28 16:34:54 - Adil Khan
Scene Analysis based on Feature Extraction and Deep Learning
Project Area of Specialization Software EngineeringProject SummaryScene classification relies on images and is essential in many computer vision applications. The scientific interest in scene classification from remotely collected images is increasing, and many datasets and algorithms are being developed. The introduction of convolutional neural networks (CNN) and other deep learning techniques contributed to vast improvements in the accuracy of image scene classification in such systems. In this project , we will apply feature extraction from various neural networks for an efficient and reliable image classification system with some applications relevant to remote sensing of scene classification.
Project Objectiveswe will research on image classification usinng CNN modern techniques , reimplement alreaady developed aggorithms developed in CNN techniques to make high accuracy , performace , efficiency and after it implement it into Android app for Practcal use. Our project bring improement in soo many areas like agriculture, AI , machine laerning, image recongnition, security , object detection, segmentation , self driving cars and robotics and many other fields .
Project Implementation MethodWe will use matlab and anaconda using paython to develop the project and aslo implement in android App.
Benefits of the ProjectHigh accuracy in image classification , performance and efficiency. our project goal is not specific , as image classification in many recent areas like,
Google Image Search
Car License Plate Recognition by camera( to detect car owner & send him Speeding ticket on his address)
Face Recognition & Iris Recognition in Mobile Security & building doors security
Real Emotion detection system detection system to analyze how they are feeling when they are hospitalized
Biometrics Recognition is the simplest example of image recognition which scans the print of thumb & remembers the print algorithim in database.
Pedestrian Brake system alert in cars to recognize the object as a person and stop the car at a safe distance thus by saving the person from a major accident.
Recognizing People images in gallery to seperate them & view only their pics( Found mostly in IOS Devices)
our improvement bring changes in All fields related to image classifation
Technical Details of Final Deliverablefirsty we reimplement the CNN model for better scene analysis integrate code in android app , upload on play store and then we will publish our research paper on image classification
Final Deliverable of the Project Software SystemCore Industry OthersOther Industries Education , IT , Medical , Agriculture , Food , Media , Security , Telecommunication Core Technology Artificial Intelligence(AI)Other Technologies Augmented & Virtual Reality, Robotics, NeuroTechSustainable Development Goals Quality Education, Decent Work and Economic Growth, Industry, Innovation and Infrastructure, Life on LandRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 29000 | |||
| ssd | Equipment | 1 | 15000 | 15000 |
| Mouse | Equipment | 1 | 500 | 500 |
| notebook for Reseach | Equipment | 3 | 1500 | 4500 |
| Matlab packages | Equipment | 5 | 1000 | 5000 |
| thesis | Miscellaneous | 1 | 4000 | 4000 |