2020-06-21 12:29:25 -0600 | commented question | Missing link for dictnet_vgg Hi @Berak, thank you for your answer, I'm looking for a text recognition CNN example like the text_end_to_end_recognitio |
2020-06-21 08:16:47 -0600 | received badge | ● Student (source) |
2020-06-21 07:05:14 -0600 | asked a question | Missing link for dictnet_vgg Missing link for dictnet_vgg Hi everyone, I was testing the OpenCV Contrib samples of the text module. Regarding to the |
2017-07-20 23:05:47 -0600 | received badge | ● Enthusiast |
2017-07-15 19:30:19 -0600 | asked a question | Segmentation of Plants on the Ground Hi, to all I want to develop a program to count plants in a crop, I have used some methods for segment the plants of the ground in OpenCV like inRange of HSV, Green Channel Segmentation and Histogram Backprojection but I have still some issues to get the total shape of the plants. I tested my algorithm with the following images in some lighting conditions For the HSV Segmentation I use four regions to get the Max and Min Values for each mask, so I sum the four masks of each segmentation to obtain an total mask, for the Green Channel Segmentation I split the channels of the BGR image and I apply the following equation to get the final mask T = 2*G - R - G, I also experimented normalizing the Green Channel to get better results in more light conditions and I apply the same equation, also I backproject the HS Histogram with a Sample of Color of a Plant and a Mask of that Plant(included in the images of the link at the bottom), so I get the following results (In this case the HSV segmentation was the best, but in the next images Normalized Green Channel have better results) My question is, is there any another method to rightly segment the plants? or how can I improve my code? I will appreciate every response, Link to Images (more) |
2017-02-20 15:50:12 -0600 | received badge | ● Scholar (source) |
2017-02-19 22:06:03 -0600 | commented answer | Problems with training a two output layer ANN_MLP from a CSV File Thanks so much, it works! I modified the csv file and I used the first Code! It works Now.... A final question, for a ten layer ANN_MLP I should add the responses as for 0 0,1,2,3,4,5,6,7,8,9 for 1 1,2,3,4,5,6,7,8,9,0 . . . for 9 9,0,1,2,3,4,5,6,7,8 etc.? |
2017-02-15 19:21:39 -0600 | received badge | ● Editor (source) |
2017-02-15 17:48:08 -0600 | asked a question | Problems with training a two output layer ANN_MLP from a CSV File Hi, I'm trying to train a ANN_MLP with two layer output, I generated a CVS file and I read it with the function TrainData::loadFromCSV() , I have the following CSV file for recognize number 1 and 0: In Opencv2 I noticed that at the final of the line de CSV files have a ;, but in the function loadFromCSV() it seams that the output argument is a one-dimmension array That's my sample code So I've got the following error: I read in this forum that:
How can I create a 2 dimmension array from TrainData::loadFromCSV() for train the ANN_MLP? I use Opencv 3.2 and GCC6 |