--- /dev/null
+# Media Vision Quality Assessment
+Media Vision has unit test(TCT or test bench) to check functionality however does not have quality check (accuracy, memory, speed) for API.
+Moreover, tizen support multiple target devices and profiles that developer hard to get compare of the various algorithms.
+MVQA(Media Vision Quality Assessment) help to measure API(convention vision algorithm or deep learning) quality across devices and profile.
+
+## Goal
+Depends on hardware target or profile, MVQA measure each API quality and show it as charts.
+
+## Prerequisite
+Dataset for testing and place it on NAS server.
+
+## How to run
+1. change current directory to script and fill nas_config.json to communicate with server.
+e.g.)
+```json
+{
+ "user": "iamroot",
+ "password": "rootpass",
+ "ip": "xxx.yyy.123.456",
+ "db_index_path": "/mvtest"
+}
+```
+note that `db_index_path` is path for dataset root.
+
+2. run python script for information
+```shell
+$ python3 mvqa_main.py -h
+```
+ or run it with test, dataset id.
+```shell
+$ python3 mvqa_main.py --dataset 1 --testset 1
+```
+
+3. program will store result as sql dataset (check mvqa.db file)
\ No newline at end of file