1.
Design a real-time and live feed system that can accurately detect and
classify multiple types of vehicles (cars, trucks, motorcycles, bicycles)
and their speed in a crowded urban environment, where lighting
conditions vary greatly and vehicles frequently overlap or occlude each
other using a double camera. Later after detecting the vehicle the model
needs to recognise the license plate and display all the content in a
dashboard. (Expected accuracy : 95%)
2. Design a real time and live feed system of custom object segmentation
to segment multiple objects. (Target class - Wild Boar) (Expected
accuracy : 95%)
Reference Link for output visualization :
https://miro.medium.com/v2/resize:fit:679/1*QKfIOzB-N1GrhULIvVxkQg.gif
3. Develop a real time and live feed system of a deep learning model to
detect cancer cells in medical images. The model would need to be able
to identify cancer cells with high accuracy, so that doctors can make
informed decisions about treatment. (Accuracy : 95%)
4. Identify different types of damage caused by a disaster. For example, we
can use detection & segmentation to identify buildings that have been
damaged by flood, roads that have been damaged by earthquakes, and
infrastructure that has been damaged by wildfires. (Accuracy : 95%)
5. Develop a deep learning image classification model that can diagnose
diseases in real time from medical images such as X-rays, MRIs, and CT
scans. (Accuracy : 95%)
6. Develop a real time design using Tensorflow Lite model to detect and
classify dust and crack in a solar panel. (Accuracy : 95%)
7. Develop an image classification model that can identify different types of
defects in manufactured products. Which could be used to improve
quality control and to reduce the number of defective products that are
shipped to customers. (Accuracy : 95%)
8. Design an action recognition system including detection system which
can detect unusual activities like fighting, chain snatching, theft and
Person with suspicious weapons. The system should be designed in such
a way that it can be integrated with traffic cams at signals and from
there it needs to be monitored. (Accuracy : 95%)
9. Design a real time, live feed based system which can detect the speed of
the cricket ball during the time of play. The detected output needs to be
displayed over the dashboard. (Accuracy : 95%)
10.Create a talking chatbot which can talk to humans like a human. It has to
build a healthy conversation with humans. It needs to intimate the
person regarding the required medication in a timely manner. It has to
be intelligent like a chatGPT based language model. The bot must surf
the internet and answer the questions if required. (It must answer 95%
of questions while final testing)
11.Create a deep learning model for voice recognition. The system must be
designed in such a way that it can recognize who is talking using their
voice and display the name of the person. (Train for 20 people minimum
and Accuracy expected is 95%)
