i am doing research in image processing. I want to use latest classification technique in my work. what is the latest classification techniques using researcher now. please list out some latest classification techniques...
Dear Anbarasa,
This is a broad question which can be narrowed by indicating the type of the image to be classified and its characteristics. If the image is a remote sensing image than depending on the problem and the required spectral, spatial and may be temporal resolution we can specify the classification methodology. Example if the problem is extracting urban features from very high spatial resolution image than it is better to go with OO classification instead of other traditional methods. On the contrary, if the problem is vegetation cover classification I would go with the pixel based methods and high to medium resolution such as SPOT and Landsat or IRS. In that case I would use clustering with unsupervised and evolutionary combined methods and techniques.
Please check my publications for more information.
If the problem is to identify anomalies in MRI images or other images than you can use deformable contours or snakes.
As far as my experience goes, the most well performing algorithms for classification are still deep neural networks. However, due to their complexity not everyone is happy to go and use them. An alternative way could be the less complex learning methods like integral channel features and aggregate channel features and all derivative models from that.
This answer mainly applies to object detection/classification.
Dear,
Latest classification techniques are
Support vector machine
Adaboosting
Decision trees
Rule-based induction
Neural networks
Memory(Case) based reasoning
Genetic algorithms
Bayesian networks
Dear,
Latest classification techniques are
Support vector machine
Adaboosting
Decision trees
Rule-based induction
Neural networks
Memory(Case) based reasoning
Genetic algorithms
Bayesian networks
Dear,
Latest classification techniques are
Support vector machine
Adaboosting
Decision trees
Rule-based induction
Neural networks
Memory(Case) based reasoning
Genetic algorithms
Bayesian networks
Dear,
Latest classification techniques are
Support vector machine
Adaboosting
Decision trees
Rule-based induction
Neural networks
Memory(Case) based reasoning
Genetic algorithms
Bayesian networks
I Have the same opinion as prof. AwD has suggested. Depending on the problem one can choose best classifier
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing