Sim Sanghyun

1. Web Fuzzer
   Concept
   Find Web vulnerability automatically

Problem
- How do I generate payload(attack code) ?
( generate payload have 2 ways 1. dumb fuzzing, 2. smart fuzzing I seleted 1, so I generate random string based on sql, html, javascript code )
- How do I judgement attack succeed ?
( If attack succeed, response code is different so I judgement using http response code )

2. Make caricature based on Computer vision (Android)
   Concept
   Make caricature look like user using Computer vision

   Problem

   • How do I detect face, nose, lip, eyes, ears
     ( This problem is very simple, we are using OpenCV, but learning model is not accurate so we can get resource using face rate )
   • How do I get index of similarity
     ( First we using OpenCV method but is not similarity, so we using Feature matching algorithm and we can improve index of similarity )
   • How can I application more faster
     ( our application is very slow because, we using many resource when find face, nose, ears, eyes and image compare. so our first solution parallel transaction. and second solution is only detect eyes and then the rest parts get using face rate)

3. Detect Ransomware based on Deep Learning
   ( developed in hackaton and our team get the grand prize (Korea Internet & Security Agency) )
   Concept
   Detect Ransomware using learning cpu, ram, network amount used pattern model in VM environment

4. One-touch insure register using Deep Learning & Blockchain (Ethereum)
   Concept
   First out application calculate insure price through personal attribute & One-touch payment based on Smart Contract