Skip to content

Releases: OlafenwaMoses/ImageAI

Test Assets

20 Dec 21:17
b286d6e
Compare
Choose a tag to compare

Resources for testing

ImageAI - PyTorch pretrained models

10 Nov 19:41
fe2d6ba
Compare
Choose a tag to compare
Merge pull request #647 from OlafenwaMoses/dependabot/pip/pillow-8.1.1

Bump pillow from 7.0.0 to 8.1.1

TF2.x Models [ Exclusives ]

04 Jan 19:24
1262edc
Compare
Choose a tag to compare
Merge pull request #448 from noobshow/patch-1

Corrected name, perhaps a typo

ImageAI 2.1.0

30 Jul 13:03
Compare
Choose a tag to compare

This is a major ImageAI release that provides a wider range of APIs, specifically for training and detecting with custom YOLOv3 models on custom datasets.

What's new

  • Training of custom YOLOv3 models on custom image datasets annotated in Pascal VOC format
  • Single and multi-model mAP evaluation of saved custom models
  • Object detection in images using custom YOLOv3 models
  • Video Object detection in using custom YOLOv3 models
  • Video Detection analysis in using custom YOLOv3 models
  • Support for file and Numpy array inputs/outputs for all custom detection

ImageAI 2.1 Essentials

30 Jul 08:27
99eb9ab
Compare
Choose a tag to compare

This release contain supporting files for training custom object detection models and performing detection. The files in this release are:

  • pretrained-yolov3.h5 : A pre-trained YOLOv3 model for transfer learning when training new detection models

  • hololens.zip : A sample detection dataset of the Hololens with Pascal VOC annotation

  • headset.zip : A sample detection dataset of Hololens and Oculus headsets with Pascal VOC annotation

  • hololens-ex-60--loss-2.76.h5 : A YOLOv3 model trained with ImageAI on the Hololens dataset

  • detection_config.json : The configuration JSON file for performing detection in images and video using the trained YOLOv3 model for Hololens.

ImageAI v2.0.3

13 Jun 13:13
afc6307
Compare
Choose a tag to compare

This is a major ImageAI release that provides a wider range of APIs for custom models and recognition. This wheel is provided for Python 3.x .

What's new:

  • Options to set video detection timeout in seconds
  • Options to save trained custom model in full
  • Support for running custom prediction with fully saved model without specifying network type
  • Support for running custom prediction with any fully saved Keras model
  • Support for converting custom trained models to Tensorflow (.pb) format
  • Support for continuous training from previously saved custom model
  • Support for transfer learning from pre-trained models for small datasets
  • Only models with increased accuracy will be saved during training
  • Support for loading and prediction with multiple custom models
  • Support for exporting models to DeepStack format

ImageAI Example Models

13 Jun 10:45
afc6307
Compare
Choose a tag to compare

In this release are models used in the sample codes.

ImageAI 2.0.2

15 Jul 21:36
fdd1944
Compare
Choose a tag to compare

This is a major ImageAI release that provides a wider range of APIs . This wheel is provided for Python 3.x .

What's new:

  • Bug Fixes

  • Option to state image size during custom image prediction model trainings

  • Object Detection and Video Object detection now returns bounding box coordinates ('box points') (x1,y1,x2, y2) for each object detected in addition to object's 'name' and 'percentage probability'

  • Options to hide 'percentage probability' and/or object 'name' from being shown in detected image or video - Support for video object detection on video live stream from device camera, connected camera and IP camera

  • Support for YOLOv3 and TinyYOLOv3 for all object detection and video object detection tasks. - Video object detection for all input types (video file and camera) now allows defining custom functions to execute after each frame, each second and each minute of the video is detected and processed. Also include option to specify custom function at once video is fully detected and processed

  • For each custom function specified, ImageAI returns the frame/seconds/minute/full video analysis of the detections that include the objects' details ( name , percentage probability, box_points), number of instance of each unique object detected (counts) and overall average count of the number of instance of each unique object detected in the case of second / minute / full video analysis

  • Options to return detected frame at every frame, second or minute processed as a Numpy array.

ImageAI 2.0.1

24 May 16:44
c7a19e3
Compare
Choose a tag to compare

This is a major ImageAI release that provides a wider range of APIs and the base implementation of future updates. This wheel is provided for Python 3.x .


What's new:

  • Addition of Custom Image Prediction model trainings using SqueezeNet, ResNet50, InceptionV3 and DenseNet121
  • Addition of Custom Image Prediction with custom trained models and generated model class json
  • Preview Release: Addition of Video Object Detection and Video Custom Objects Detection (object tracking)
  • Addition of file, numpy array and stream input types for all image prediction and object detection tasks (file inputs only for video detections)
  • Addition of file and numpy array output types for object detection and custom object detection in images
  • Introduction of 4 speed modes ('normal', 'fast', 'faster' and 'fastest') for image prediction, allowing prediction time to reduce by 50% at 'fastest' while maintaining prediction accuracy
  • Introduction of 5 speed modes ('normal', 'fast', 'faster', 'fastest' and 'flash') for all object detection and video object detection tasks, allowing detection time to reduce by over 80%

    with 'flash' detection accuracy balanced with 'minimum_percentage_probability' which is kept at low values
  • Introduction of rate of frame detections, allowing developers to adjust intervals of detections in videos in favour of real time/close to real time result.


IdenProf Dataset Pre-Trained Models

24 May 23:50
5ebcb87
Compare
Choose a tag to compare

This release contains the models pre-trained on IdenProf, the Identifiable Professionals datatset as well as the model JSON class mapping file.