33 votes
Accepted

What is the difference between semantic segmentation, object detection and instance segmentation?

Object Detection : is the technology that is related to computer vision and image processing. It's aim? detect objects in an image. Semantic Segmentation : is a technique that detects , for each pixel ...
  • 1,994
31 votes
Accepted

What does the notation mAP@[.5:.95] mean?

mAP@[.5:.95](someone denoted mAP@[.5,.95]) means average mAP over different IoU thresholds, from 0.5 to 0.95, step 0.05 (0.5, 0....
  • 4,096
25 votes
Accepted

How to calculate mAP for detection task for the PASCAL VOC Challenge?

To answer your questions: Yes your approach is right Of A, B and C the right answer is B. The explanation is the following: In order to calculate Mean Average Precision (mAP) in the context of ...
  • 2,166
24 votes

What is the difference between Inception v2 and Inception v3?

In the paper Batch Normalization,Sergey et al,2015. proposed Inception-v1 architecture which is a variant of the GoogleNet in the paper Going deeper with convolutions, and in the meanwhile they ...
24 votes
Accepted

Why convolutions always use odd-numbers as filter size

The convolution operation, simply put, is combination of element-wise product of two matrices. So long as these two matrices agree in dimensions, there shouldn't be a problem, and so I can understand ...
18 votes
Accepted

number of parameters for convolution layers

Actually it's $49C*C$, the first $C$ is the number of input channels, and the second $C$ is the number of filters. Quote from CS231n: To summarize, the Conv Layer: Accepts a volume of size ...
  • 4,096
15 votes

How to calculate mAP for detection task for the PASCAL VOC Challenge?

There is a nice and detailed explanation with an easy to use code on my Github. Certainly it will help you guys.
11 votes
Accepted

What is fractionally-strided convolution layer?

Here is an animation of fractionally-strided convolution (from this github project): where the dashed white cells are zero rows/columns padded between the input cells (blue). These animations are ...
  • 8,667
9 votes
Accepted

Optimizer for Convolutional neural network

Yes, you can use the same optimizers you are familiar with for CNNs. I don't think that there is a best optimizer for CNNs. The most popular in my opinion is Adam. However some people like to use a ...
  • 422
8 votes

How to deal with large training data?

I do something similar with keras and GPU training, where i also have only a small memory amount available. The idea would be split the numpy files into smaller ones, let's say 64 samples per file and ...
  • 2,452
8 votes
Accepted

Using Neural Networks to extract multiple parameters from images

A CNN could be a good choice for this task if you expect variation in the original image scale, rotation lighting etc, and also have a lot of training data. The usual CNN architecture is to have ...
  • 27.6k
8 votes
Accepted

How to implement global contrast normalization in python?

there are multiple issues with the code: You force the values in the image to be uint8 (8-bit integer). Since the values are floats they will be casted/rounded to either 0 or 1. This will later be ...
  • 96
8 votes
Accepted

What is the difference between Dilated Convolution and Deconvolution?

In sort of mechanistic/pictorial/image-based terms: Dilation: ### SEE COMMENTS, WORKING ON CORRECTING THIS SECTION Dilation is largely the same as run-of-the-mill convolution (frankly so is ...
8 votes
Accepted

What is the meaning of hand crafted features in computer vision problems?

"Hand Crafted" features refer to properties derived using various algorithms using the information present in the image itself. For example, two simple features that can be extracted from images are ...
  • 166
8 votes

What is difference between Fully Connected layer and Bilinear layer in CNN?

I quote the answers from What is a bilinear tensor layer (in contrast to a standard linear neural network layer) or how can I imagine it?. A bilinear function is a function of two inputs $x$ and $y$ ...
  • 13.5k
8 votes
Accepted

Does image's background matter for detector training (CNN)?

Of course, it matters. Which one is best completely depends on your problem. The golden rule for machine learning problems is that you want the data you train on to be as representative as the data ...
8 votes
Accepted

How to make two parallel convolutional neural networks in Keras?

You essentially need a multi-input model. This can only be done through keras' functional api and can work with the pretrained nets in keras.applications. To create ...
  • 7,618
8 votes

How can you build a model that extracts data out from receipts?

The simplest pipeline would be to do the following: OCR Named Entity Extraction Entity Disambiguation OCR This is basically transforming your receipts into plain text. If you have scans (pictures) ...
7 votes
Accepted

Convnet training error does not decrease

In your convnet code, you compute the cross entropy manually: ...
  • 1,872
7 votes

What is the difference between Dilated Convolution and Deconvolution?

Though both seem to be doing the same thing, which is up-sampling a layer, there's a clear margin between them. First we talk about Dilated Convolution I found this nice blog on above topic. So as ...
7 votes
Accepted

Data preprocessing: Should we normalise images pixel-wise?

We first center our data by subtracting the mean of the batch. We also divide by the standard deviation, so our formula becomes: $ z = \frac{x - \mu}{\sigma} $ where: $ x $ is the pixel value $ \...
  • 2,462
7 votes
Accepted

Is There any RNN method used for Object detection

Yes, there have been many attempts, but perhaps the most noteable one is the approach described in the paper of Andrej Karpathy and Li Fei-Fei where they connect a CNN and RNN in series (CNN over ...
7 votes

Why choose TensorFlow?

I will outline some points about the libraries and point you to some good comparisons that I have read. The GitHub star counts are just another reference point to help compare popularity. While I don'...
  • 14.1k
7 votes

mAP scores on tensorboard (Tensorflow Object Detection API) are all 0 even though the loss value is low

I recommend several checks to make sure you get reasonable mAP@IoU scores for object detection API: Try varying the Intersection over Union (IoU) threshold, e.g 0.2-0.5 and see if you get an increase ...
6 votes
Accepted

What is the depth of an image in Convolutional Neural Network?

It means that the number of filters (a.k.a. kernels, or feature detector) in the previous convolutional layer is 96. You may want to watch the video of the lecture, and in particular this slide, ...
6 votes

What does the notation mAP@[.5:.95] mean?

You already have the answer from Icyblade. However, I want to point out that your Average Precision formula is wrong. The formula $\frac{\#TP(c)}{\#TP(c) + \#FP(c)}$ is the definition of precision, ...
  • 61
6 votes

Why do we need for Shortcut Connections to build Residual Networks?

The short answer is that when a net is very deep it becomes very difficult for gradients to propagate backwards all the way. Skip connections offer "short cuts" for gradients to propagate further and ...
  • 161
6 votes
Accepted

The goal of fine tuning

Fine tuning means changing the weights such that the VGGNet can perform the task you want in your dataset. The reason why fine-tuning is not called training (which is what you are doing) is because it ...
  • 5,714
6 votes
Accepted

Which is the fastest image pretrained model?

The answer will depend on some things such as your hardware and the image you process. Additional, we should distinguish if you are talking about a single run through the network in training mode or ...
  • 14.1k
6 votes

How to detect blocks of texts in document images

I would approach the text block amalgamation as a clustering problem. If you define a suitable distance metric or a neighbour predicate between the individual text boxes, you could group the boxes and ...

Only top scored, non community-wiki answers of a minimum length are eligible