We often want to generally go to mess at some time when it is less crowdy so we can get sitting easily. Sometimes we even want to order something special from the mess during it's canteen time. So how to do do these things from your mobile phone ?

Developed an app for accessing the crowd status in IIT Jammu mess, analyzed the video to get the count of members utilizing mess at a particular time. This app provides real-time information about the mess. User can book some special food (if possible) for the guest.

The Mess Application was built using android studio . A registeration page for a new user for registeration followed by user login to authenticate the user . A Dynamic menu(which shows the Breakfast , lunch and dinner of current day) which keeps changing according to calender, mess menu of whole week was attached, in addition to it contact support for the user was also provided . User can place orders on the required date and time. Firebase database was used as backend database which keeps track of real time data access and authentication . To implement counting of number of people in mess we used firebase’s ML-developer kit which helped us to count the number of people currently in mess.

Concerned About security of data in your phone. As we know securing once data is of huge importance and phone may fall into wrong hands. How to secure data in your mobile phone from others? This App is aimed to achieve highest security in authenticating users.

Developed an android app with three levels of login consisting a text based, image segmentation based and graphical password. Google’s reCaptcha API is used to further reduce chances of bot attack.

Features :

• Users would be given a registration form that has to be filled with required details.
• Next users would be asked to set password for first level, second level and third level subsequently.
• After the passwords are set for the three level users can now login into the system.
• While login the system will ask for the first level password. On entering correct password, second level password is asked and then third one.
• After the user has provided correct password in the third level, he gets authenticated and can now access the system.

Levels of authentication :

• First Level: The first level is a conventional password system i.e. text based password or a password phrase. Users would have to set a text password initially based on some specifications.
• Second Level: The second level is a graphical password method where users have to set password based on some color combinations through RGB button combinations.
• Third Level: The third level is an image based password where users can upload their desired image into the system and then create password by segmenting it and assigning them serial numbers.
• Additional layer of authentication has been provided at the end using google's reCaptcha API(Humans v/s Bots).

Worried about your message being read by another user. How can we make sure that even if a message is read by a user for whom it was not intended, cannot understand the message?

Developed an android app to prevent man in the middle attack. The user can send the encrypted message, as SMS, using this app and the person receiving the SMS can decrypt through the Secret Key using this app.

Features :

• Encrypt: The user can use this feature to encrpyt the message and send it to required person via SMS .
• Decrypt: The user is allowed to view the complete message which is decrypted as he selects the messages from the Inbox
• Fast and Easy to use , No Internet Needed , It is highly reliable and secure.

A Python code for recogntion of a person via face detection and recognition. We trained model on our training data and when a testing image comes or a frame is extracted from the live stream , the model matches the two images on basis of simlarity by considering 68 points on the face . Model used classifer which uses SVM(support vector machine) . Model uses dlib , face_recognition and opencv library in our model . In order to improve the accuracy , we tweeked several parameters like changing tolerance of our model etc .The model makes a square box on the faces of the people detected and if the person is known then it recoganises it and displays the name else it displays unknown.

Features :

• The system detects and recognises faces of people in live stream.
• It has quite high accuracy and we achieved this by tuning with other parameters such as tolerance etc.
• It is integrated well with hardware like raspberry pie.

Applications :

• We may deploy the system at the entrance of the class/institute/hostel. Every authorized member is required to get access through the system.We may also choose an audio to announce an un-authorized access and alarm the security centre.
• We may deploy this model for automated attendance system.

Want to Sniff the packets travelling in the network?

Created a Web GUI using eel library to interact between Python backend and frontend from HTML, CSS and Javascript.

Features :

• Is able to show all the headers till transport layer for packets received in real time. You have to see whether TCP or UDP is used at Transport layer from IP header information. In case of ICMP it shows headers till ICMP.
• Program is able to show the application layer data. Assumpting that we are only taking about application layer protocols that do not encrypt data. HTTP, DNS etc.
• The program provides regex matching on application layer data. For example program can tell if the packet application layer payload has a "password" keyword in it or not.