Semantic-Segmentation

Implemented Semantic Segmentation in PyTorch and Tensorflow.

Dataset

The dataset has been generated as part of the lyft challenge

Sample input and output of the image

What it does?

The aim of the project is to classify each pixel in the image into one among the classes mentioned in the table below:

Value	Tag	Description
0	Unlabeled	Elements that have not been categorized are considered Unlabeled. This category is meant to be empty or at least contain elements with no collisions.
1	Building	Buildings like houses, skyscrapers,... and the elements attached to them. E.g. air conditioners, scaffolding, awning or ladders and much more.
2	Fence	Barriers, railing, or other upright structures. Basically wood or wire assemblies that enclose an area of ground.
3	Other	Everything that does not belong to any other category.
4	Pedestrian	Humans that walk or ride/drive any kind of vehicle or mobility system. E.g. bicycles or scooters, skateboards, horses, roller-blades, wheel-chairs, etc.
5	Pole	Small mainly vertically oriented pole. If the pole has a horizontal part (often for traffic light poles) this is also considered pole. E.g. sign pole, traffic light poles.
6	RoadLine	The markings on the road.
7	Road	Part of ground on which cars usually drive. E.g. lanes in any directions, and streets.
8	SideWalk	Part of ground designated for pedestrians or cyclists. Delimited from the road by some obstacle (such as curbs or poles), not only by markings. This label includes a possibly delimiting curb, traffic islands (the walkable part), and pedestrian zones.
9	Vegetation	Trees, hedges, all kinds of vertical vegetation. Ground-level vegetation is considered Terrain.
10	Vehicles	Cars, vans, trucks, motorcycles, bikes, buses, trains.
11	Wall	Individual standing walls. Not part of a building.
12	TrafficSign	Signs installed by the state/city authority, usually for traffic regulation. This category does not include the poles where signs are attached to. E.g. traffic- signs, parking signs, direction signs...
13	Sky	Open sky. Includes clouds and the sun.
14	Ground	Any horizontal ground-level structures that does not match any other category. For example areas shared by vehicles and pedestrians, or flat roundabouts delimited from the road by a curb.
15	Bridge	Only the structure of the bridge. Fences, people, vehicles, an other elements on top of it are labeled separately.
16	RailTrack	All kind of rail tracks that are non-drivable by cars. E.g. subway and train rail tracks.
17	GuardRail	All types of guard rails/crash barriers.
18	TrafficLight	Traffic light boxes without their poles.
19	Static	Elements in the scene and props that are immovable. E.g. fire hydrants, fixed benches, fountains, bus stops, etc.
20	Dynamic	Elements whose position is susceptible to change over time. E.g. Movable trash bins, buggies, bags, wheelchairs, animals, etc.
21	Water	Horizontal water surfaces. E.g. Lakes, sea, rivers.
22	Terrain	Grass, ground-level vegetation, soil or sand. These areas are not meant to be driven on. This label includes a possibly delimiting curb.

Architecture

This is implemented based on the research paper U-Net: Convolutional Networks

The summary of the architecture as presented in the paper

Predictions