출처 : https://github.com/abreheret/PixelAnnotationTool
이 툴을 간단히 소개하자면, semantic segmentation에 이미지를 학습시키려면 object별로 pixel을 구분하면 구분은 color 및 label등으로 한다. 이 툴을 통하면 간단히? object 구분하는 label mask 이미지를 만들수 있고, my github에 Semantic-Segmentation에서 자신이 구현하고자 하는 Segmentation을 모델링 할 수 있다
- 1)select the config(json) you want to used
- 2)colour the object with the selected colour; you could zoom in/out by changing value of Resize factor and change size of brush by changing Circle size
- 3)if ther are multiple object then select different colour for each object and colour the object a bit
- 4)colour background with different colour
- 5)check watershed mask and click on watershed
- 6)you could change value of alpha mask to make mask more visible
- 7)fine tune mask a bit if required
- 8)see final mask clearly by setting alpha mask 1.0
- 9)save image
- 10)do it for all images in image
참고 유튜브 (https://www.youtube.com/watch?v=tX-xcg5wY4U)
사전에 하기와 같이 label과 rgb를 정의해놓으면 편할 것 같다. name,r,g,b Animal,64,128,64 Archway,192,0,128 Bicyclist,0,128, 192 Bridge,0, 128, 64 Building,128, 0, 0 Car,64, 0, 128 CartLuggagePram,64, 0, 192 Child,192, 128, 64 Column_Pole,192, 192, 128 Fence,64, 64, 128 LaneMkgsDriv,128, 0, 192 LaneMkgsNonDriv,192, 0, 64 Misc_Text,128, 128, 64 MotorcycleScooter,192, 0, 192 OtherMoving,128, 64, 64 ParkingBlock,64, 192, 128 Pedestrian,64, 64, 0 Road,128, 64, 128 RoadShoulder,128, 128, 192 Sidewalk,0, 0, 192 SignSymbol,192, 128, 128 Sky,128, 128, 128 SUVPickupTruck,64, 128,192 TrafficCone,0, 0, 64 TrafficLight,0, 64, 64 Train,192, 64, 128 Tree,128, 128, 0 Truck_Bus,192, 128, 192 Tunnel,64, 0, 64 VegetationMisc,192, 192, 0 Void,0, 0, 0 Wall,64, 192, 0