Image Processing - Revision history

Florian at 10:20, 23 September 2017

2017-09-23T10:20:09Z

2017-09-23T10:19:50Z

2017-09-23T09:51:02Z

2017-09-23T09:49:13Z

Description of Neural Network Image Processing approach

2017-08-19T07:32:59Z

Added paper on edge detection (reference to start with)

New page

A general paper discussing EDGE detection by using neural networks:

<pdf>File:Edge-detection-using-convolutional-neural-network.pdf</pdf>

← Older revision		Revision as of 10:20, 23 September 2017
Line 1:		Line 1:
−	A research group from the university of California used Neuronal Networks for EDGE detection: [https://github.com/s9xie/hed github repository].	+	A research group from the university of California used Neuronal Networks for EDGE detection: [https://github.com/s9xie/hed github repository], see paper below.

	They used [http://caffe.berkeleyvision.org/ Caffe Deep learning framework]. The same concepts can be applied to our case, maybe using [https://caffe2.ai/ Caffe2]. NOTE: the alternative would be TensorFlow, but the mobile deployment is too heavy for the old raspberry PI and needs a huge SD card, see [https://github.com/samjabrahams/tensorflow-on-raspberry-pi/blob/master/GUIDE.md here], meanwhile it should be easier with Caffe2, see [https://caffe2.ai/docs/getting-started.html?platform=raspbian&configuration=compile here].		They used [http://caffe.berkeleyvision.org/ Caffe Deep learning framework]. The same concepts can be applied to our case, maybe using [https://caffe2.ai/ Caffe2]. NOTE: the alternative would be TensorFlow, but the mobile deployment is too heavy for the old raspberry PI and needs a huge SD card, see [https://github.com/samjabrahams/tensorflow-on-raspberry-pi/blob/master/GUIDE.md here], meanwhile it should be easier with Caffe2, see [https://caffe2.ai/docs/getting-started.html?platform=raspbian&configuration=compile here].

← Older revision		Revision as of 10:19, 23 September 2017
Line 6:		Line 6:
	So basically we need an Image to Image neural network that recognizes the parts of the agricultural fields black and the rest white (or the opposite).		So basically we need an Image to Image neural network that recognizes the parts of the agricultural fields black and the rest white (or the opposite).
	To simplify the training of the network it could be a good idea to develop an iterative image recognition algorithm, based on color, texture, ... first and use it to provide learning data with google maps. The automatically elaborated pictures can then just be corrected with human interaction (smooth workflow needed!) to provide the needed gain in quality.		To simplify the training of the network it could be a good idea to develop an iterative image recognition algorithm, based on color, texture, ... first and use it to provide learning data with google maps. The automatically elaborated pictures can then just be corrected with human interaction (smooth workflow needed!) to provide the needed gain in quality.
		+
		+	<pdf>File:holistically-nested-edge-detection-2015.pdf</pdf>

@@ Line 1: / Line 1: @@
 A research group from the university of California used Neuronal Networks for EDGE detection: [https://github.com/s9xie/hed github repository].
 They used [http://caffe.berkeleyvision.org/ Caffe Deep learning framework]. The same concepts can be applied to our case, maybe using [https://caffe2.ai/ Caffe2].
 So basically we need an Image to Image neural network that recognizes the parts of the agricultural fields black and the rest white (or the opposite).
 To simplify the training of the network it could be a good idea to develop an iterative image recognition algorithm, based on color, texture, ... first and use it to provide learning data with google maps. The automatically elaborated pictures can then just be corrected with human interaction (smooth workflow needed!) to provide the needed gain in quality.

← Older revision		Revision as of 09:49, 23 September 2017
Line 1:		Line 1:
−	A ~~general paper discussing~~ EDGE detection by using ~~neural networks~~:	+	A research group from the university of California used Neuronal Networks for EDGE detection: [https://github.com/s9xie/hed github repository].
		+	They used [http://caffe.berkeleyvision.org/ Caffe Deep learning framework]. The same concepts can be applied to our case, maybe using [https://caffe2.ai/ Caffe2].

−	~~<pdf>File:Edge-detection-using-convolutional-~~neural-network.~~pdf</pdf>~~	+
		+	So basically we need an Image to Image neural network that recognizes the parts of the agricultural fields black and the rest white (or the opposite).
		+	To simplify the training of the network it could be a good idea to develop an iterative image recognition algorithm, based on color, texture, ... first and use it to provide learning data with google maps. The automatically elaborated pictures can then just be corrected with human interaction (smooth workflow needed!) to provide the needed gain in quality.