Questions tagged [object-detection]

Object detection is a computer-vision and image-processing technique for locating instances of objects in images or videos. Common applications include face detection and object tracking. Object detection algorithms typically leverage machine learning or deep learning to produce meaningful results.

198 questions with no upvoted or accepted answers
Filter by
Sorted by
Tagged with
1
vote
0answers
307 views

YOLO v3 complete architecture

I am attempting to implement YOLO v3 in Tensorflow-Keras from scratch, with the aim of training my own model on a custom dataset. By that, I mean without using pretrained weights. I have gone through ...
1
vote
1answer
360 views

Detecting address labels using Tensorflow Object Detection API

I am experimenting with the Tensorflow Object Detection API on a Windows 7 machine. I am trying to detect US address labels (and similar blocks of text) as they appear on a piece of mail or an ...
1
vote
0answers
20 views

Is it beneficial to train 1 object detector for N classes vs N object detectors each for a different class?

I always wondered if training an object detector to recognize (let's say), dogs and cats performs better or worse than training 2 object detectors, one for cats only and one for dogs only. Would this ...
1
vote
0answers
35 views

Keras - How to using Cnn trained model on video image.?

I have trained Cat and dog image using CNN with Python ( total 10K, 8K for training and 2K for testing ). I want to make prediction in a video. Video Contain both cat and dog at a single moment. How ...
1
vote
0answers
56 views

Using tensorflow object detection in another model

I am trying to use Tensorflow (tf) object detection API models in another custom model I built. Specifically, I am trying to do: jointly train tf object detection models Y with another model X. in a ...
1
vote
0answers
22 views

Does anyone have suggestions for how Goodfellow et al detected digits from the larger Street View images in their 2014 paper?

In their 2014 paper titled, "Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks", they write, We use an automated method (beyond the scope of this ...
1
vote
1answer
588 views

SSD based on ResNet-101 doesn't improve over SSD-VGGNet

I am training a SSD model for detecting mobile cranes. The training dataset contains 1,000 images and test set over 400 images. About 200 epochs gave mAP 83%, but my target is 90%. So I trained SSD-...
1
vote
0answers
87 views

Add training data to YOLO post-training

I've been playing around with YOLOv3 and obtaining some good results on the ~20 custom classes I trained. However, one or two classes look like they can use some additional training data (not a lot, ...
1
vote
0answers
75 views

Training a convoltion neural network for localization

I am new into convolutional neural networks and I am trying to build a cnn for localization of object on tensorflow. Currently I am trying to replicate a cnn similar to that of Alexnet and then want ...
1
vote
0answers
66 views

Better to crop or compress training data?

I've been trying to make an object recogniser in Tensorflow and have used labelImg to classify large electrical transmission towers at varying distances. In order to make 10-16MP (~2-7MB) images train ...
1
vote
0answers
1k views

How to perform Instance Segmentation using Tensorflow?

I used Tensorflow Object Detection API for a custom dataset based on the instructions at this help document.As required , collected the dataset,annotated it in PASCAL VOC XML format,split into ...
0
votes
0answers
6 views

Reduce Training steps for SSD-300

I am new to deep learning and I am trying to train my SSD-300 (single shot detector) model which is taking too long. For example even though I ran 50 epochs, it is training for 108370+ global steps. I ...
0
votes
0answers
5 views

Soft-NMS during Inference

I have read the paper about the soft-NMS and the problem that tries to solve. Does it make sense to use this algorithm during inference apart from training? It smooths the predictions on the bounding ...
0
votes
0answers
26 views

Does having two different models improve performance for underrepresented classes?

I am currently working on a dataset that has approximately 7000 annotations, but suffers from severe class imbalance (there are 1331 annotations for the most represented class, and 77 for the least ...
0
votes
0answers
33 views

Fine-tuning a pre-trained Tensorflow model

I have two questions regarding fine-tuning a pre-trained Tensorflow model on a mix of high-quality and low-quality images.In other words, ' Does the large contrast in image quality confuses the model ...
0
votes
0answers
13 views

How can I detect symbols/icons on a label?

Given this label: How can I identify that the CE mark is there? I've tried template matching with OpenCV but results are quite inconsistent because of symbol-template size differences (scaling & ...
0
votes
0answers
9 views

What kinds of changes can I attempt on my object detector .config file to improve the detection accuracy?

I have trained an object detection model with 2 classes, around 7500 images, and approx. 10,000 annotations per class. I was able to fine-tune Faster R-CNN with ResNet (V1) from the Tensorflow Object ...
0
votes
0answers
6 views

Why do CNN regression models fail to localise an object when the scale, aspect & image type has significant variance?

Background I'm wanting to understand how to build an object detection algorithm from scratch. My initial thoughts were that architectures like YOLO and Faster-RCNN would find any object given enough ...
0
votes
0answers
19 views

Why is VGG16 better for object localisation than MobileNet?

Prologue I have been trying to perform object localization to provide the [x1,y1,x2,y2] coordinates of objects in an image using Keras. I was stuck for ever because I was using MobileNetV2 as my ...
0
votes
0answers
30 views

Extract features using Bounding Box

I have a ground truth bounding box for a 3d object. I would like to extract useful features for the object. My goal is to concatenate these visual object features with language features (from the ...
0
votes
0answers
22 views

COVID-19 Rapid Test Result Image Detection

I fine-tuned an Inception V3 model provided in AWS SageMaker to detect COVID-19 Rapid Test Results (see the image below for an example). I provided about 20 pictures of negative and about 20 pictures ...
0
votes
0answers
25 views

Minibatch SGD performs better than Adam for Region proposal network training

I am using both minibatch SGD (with momentum) and Adam for training a region proposal network. The library used is KERAS. The batch size in both cases is 5 and initial learning rate is 0.01. The ...
0
votes
0answers
4 views

How to create COCO format data out of list of boxes

I have $N$ images. I have a script that extracts boundary boxes of an object that I am interested in. For each image, I may get $m$ boxes. There is only one item that I am interested in which is cat. ...
0
votes
0answers
10 views

What does the Region Proposal Network output in Faster-RCNNs?

Does it output corrections and offsets to the anchor boxes(that were generated by using some specific aspect ratios and scales)? Also if this the answer is YES, Suppose I have 3 scales - [8,16,32] and ...
0
votes
0answers
17 views

how to make image classification model to detect object and track object in a video

i have built a disaster classification model . now i want to use this classification model to detect ex- cyclone in a video to draw bounding box around it and track the cyclone if possible.is it ...
0
votes
0answers
13 views

can i do object detection on pretrained classification model...?

so i built a natural disaster classification model using transfer learning Renet50(tensor flow) got 98% accuracy and now instead of just classifying natural disaster lets say a cyclone appeared in ...
0
votes
0answers
10 views

Compute IoU for each class in Mask R-cnn

I'm trying to compute the IoU, with the matterport Mask R-cnn implementation, for each class (13 in total) that i have in my dataset. For now i managed to compute the average IoU for all the classes ...
0
votes
0answers
16 views

How to extract undetected object class labels after extracting `y_pred` from `predictions. pth` inference file for Mask rcnn

I training maskrcnn on a custom dataset with two classes (1 and 2). After testing, I get some files segm.json, predictions.pth, <...
0
votes
0answers
18 views

Extracting features from bounding boxes of CornerNet

I am using the CornerNet model. I want to extract features from specific bounding boxes that have been detected. Unlike Faster RCNN,...
0
votes
0answers
35 views

Implementing class weighting in Faster RCNN

I have a dataset (around 45,000 screenshots) of UI elements (UI trees containing element types and bounding boxes) and associated screenshots: The dataset is highly imbalanced with the button element ...
0
votes
0answers
9 views

Understanding how anchors are created in a regional proposal network

I understand that in Faster R-CNN, the image is fed into a pre-trained CNN (such as VG16). So say I have a 37x50x512 feature map. Firstly, I assume that each feature map (37x50x1) is fed into the RPN? ...
0
votes
0answers
10 views

Why do images in Open CV template matching get matched to the templates more than once?

The following code reads the set of input tiff files under tiffile and template tiff files under files and performs template matching on them. After the image gets matched, opencv contour extracts the ...
0
votes
0answers
48 views

Spatial positional encodings Vs Learned positional encodings(Object queries)

I have been trying to understand facebook's Detection transformer(DeTr) paper. Architecture Most of the explanation about the architecture is straightforward. I don't especially understand the ...
0
votes
1answer
25 views

How to feed high resolution images to the model?

I want to do object detection , normally we have a size of image 256 x 256 or 128 x 128 but what if we want to feed high ...
0
votes
0answers
27 views

How to improve the performance of an object detector model?

I have few questions regarding improving the performance of my object detection model. When there is color match between person uniform and the background, it becomes difficult for my model to ...
0
votes
0answers
177 views

yolo v4 vs yolo v4-tiny which is better for real time object detection or is that any that is better for real time object detection?

When I am comparing Yolo v4 and Yolo v4 tiny, I notice that the tiny one does significantly worse. It may be caused by the small amount of data I use for testing, so I want to ask in a normal amount ...
0
votes
0answers
20 views

Object Detection - How to evaluate saved model weights?

I've recently completed training an object detection model but only printed out the losses for each epoch. Since I have the saved model weights, how would I go about evaluating the accuracy/precision? ...
0
votes
0answers
12 views

Automated tools to generate synthetic training images out of synthetic 3D models and 2D backgrounds

I have 3D meshes and textures for a dozen of objects, which have to be detected in synthetic images. I have 2D textures of backgrounds these objects will be visible in front of. Object detection will ...
0
votes
0answers
14 views

How to annotate the object more efficiently

I want to perform object detection and object counting on the given below image using neural networks. My first step will be to annotate and label each object present in the training set of images. In ...
0
votes
0answers
12 views

Value of AP and AR are -1.000 in evaluation

As I understand, the AP amd AR calculation are as of follow: ...
0
votes
0answers
10 views

Train object detection to detect floor type

What is the best approach to train object detection to efficiently detect floor type in an image (e.g., hardwood floor, tile floor, etc.)? I don't need a full mask or bounding box, I just need to ...
0
votes
0answers
43 views

What is the difference between Salient and Generic Object Detection?

There are two types of Detection methods in general. Generic and Salient. Question: What is the exact difference in between the two? Broad Difference:
0
votes
0answers
19 views

How can I use COCOAPI/PyCOCOTools to evaluate the performance of Tensorflow Lite models

I have used the Tensorflow Object Detection API to train models on a custom dataset. The tensorflow object detection API also allows evaluating the trained models on a test set and gives results in ...
0
votes
0answers
11 views

Scalable Infrence Server for Object Detection

I have created a Django service (nginx + Gunicorn) for object detection models. For my case i have 50+ models with resnet 50 based back bone. Server Machine Specification: 16 CPU 64 GB Ram I have ...
0
votes
0answers
16 views

Can't get a regression problem to converge

I am working on implementing a really simple version of YOLO to learn about pytorch and building deep learning models. My dataset consists of images which have two MNIST digits placed somewhere on the ...
0
votes
0answers
89 views

What is the difference between a bounding box and ROI (Region of Interest)

I was reading about the Fast RCNN for object detection. From what I understand, it uses pre-computed ROI's (using selective search) and uses these to predict the bounding box offsets and uses smooth ...
0
votes
0answers
58 views

Best algorithm for real-time instance segmentation in videos?

I want to do object detection in real-time (meaning localization and classification) on videos (25 FPS), but with the added constraint that my training data is labelled using binary masks, rather than ...
0
votes
1answer
13 views

Search for implementation of Faster RCNN

What are the best written and best structured Faster RCNN implementations that you know? Please provide references.
0
votes
0answers
6 views

Multi-object detection within single image

Given an image with multiple objects within it, I would like to train a CNN to output vector of labels corresponding to the presence/absence of objects within the image. I would like to know whether ...
0
votes
0answers
6 views

Automatic Video Editing via Neural Networks; Identify the best start and end frame to trim a video?

I have an abundance of videos from high speed cameras of golf swings. With this data, my ultimate goal is to apply to to these videos an already defined human pose estimation model. However, a clear ...