tensorflow / tensorflow/models

How to improve the performance of the object detection example in the Android sample?

Open
#11,257 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

models:research type:support
Dominant language
Python
Stars
77.7k
Forks
44.8k
PR merge metrics
No merged PRs in 30d

Description

I trained ssd_mobilenet_v2 object detection model by https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/running_on_mobile_tf2.md and https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2_detection_zoo.md
and convert to tflite to run on android app
https://github.com/tensorflow/examples/tree/master/lite/examples/object_detection.
The object detection processing time is approximately 30ms above.
I have used nnapi to increase speed, I hope to achieve an object detection time of less than 30ms . What methods can I use to further improve the detection speed?

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reviewing the linked TensorFlow object-detection documentation and the lite/examples/object_detection Android sample. Profile the existing NNAPI-backed inference path to identify the remaining processing cost, then determine which supported changes could meet the requested under-30ms target and validate the result on the same device and model.

Written by the indexing model from the issue text.

Assessment

Tech stack
android, tensorflow
Domain
machine-learning, mobile-dev
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.