Skip to content

Single Shot MultiBox Detector implemented with TensorFlow

License

Notifications You must be signed in to change notification settings

arabian9ts/SSD300

Folders and files

NameName
Last commit message
Last commit date

Latest commit

995a014 · May 13, 2018
Oct 12, 2017
Mar 17, 2018
Mar 17, 2018
Oct 12, 2017
Sep 27, 2017
Nov 8, 2017
Oct 12, 2017
Mar 17, 2018
Feb 28, 2018
Oct 30, 2017
Oct 12, 2017
Mar 17, 2018

Repository files navigation

SSD300

Single Shot MultiBox Detector implemented with TensorFlow

Dependencies

python3.6.1

  • numpy
  • skimage
  • TensorFlow
  • matplotlib
  • OpenCV

Usage

  1. Import required modules
import tensorflow as tf
import numpy as np

from util.util import *
from model.SSD300 import *
  1. Load test-image
img = load_image('./test.jpg')
img = img.reshape((300, 300, 3))
  1. Start Session
with tf.Session() as sess:
        ssd = SSD300(sess)
        sess.run(tf.global_variables_initializer())
        for ep in range(EPOCH):
            ...
  1. Training or Evaluating you must just call ssd.eval() !
...

_, _, batch_loc, batch_conf, batch_loss = ssd.eval(minibatch, actual_data, is_training=True)

...

Test Training

you have to extract data-set from zip files. decompress all zip files in datasets/ and move to voc2007/ dir.

$ ls voc2007/ | wc -l    #  => 4954
$ ./setup.sh
$ python train.py

Present Circumstances

I'm checking and testing SSD model, so this model may not be complete.

If I have overlooked something, please tell me.

Welcome PullRequest or E-mail

Releases

No releases published

Packages

No packages published