From bcebe2cde225236b0f88e77a3219e3cf4a9cdb73 Mon Sep 17 00:00:00 2001 From: "Kwabena W. Agyeman" Date: Wed, 30 Oct 2019 00:24:43 -0700 Subject: [PATCH] Add mobilenet script examples Yes, mobilenet runs ont the OpenMV Cam now. It's midly interesting, however, it's really not meant to be used by itself. --- .../tf_mobilenet_search_whole_window.py | 56 +++++++++++++++++ .../tf_mobilenet_serach_just_center.py | 62 +++++++++++++++++++ 2 files changed, 118 insertions(+) diff --git a/scripts/examples/25-Machine-Learning/tf_mobilenet_search_whole_window.py b/scripts/examples/25-Machine-Learning/tf_mobilenet_search_whole_window.py index e69de29bb..8a2c559d6 100644 --- a/scripts/examples/25-Machine-Learning/tf_mobilenet_search_whole_window.py +++ b/scripts/examples/25-Machine-Learning/tf_mobilenet_search_whole_window.py @@ -0,0 +1,56 @@ +# TensorFlow Lite Mobilenet V1 Example +# +# Google's Mobilenet V1 detects 1000 classes of objects +# +# WARNING: Mobilenet is trained on ImageNet and isn't meant to classify anything +# in the real world. It's just designed to score well on the ImageNet dataset. +# This example just shows off running mobilenet on the OpenMV Cam. However, the +# default model is not really usable for anything. You have to use transfer +# learning to apply the model to a target problem by re-training the model. +# +# NOTE: This example only works on the OpenMV Cam H7 Pro (that has SDRAM) and better! +# +# In this example we slide the detector window over the image and get a list +# of activations. Note that use a CNN with a sliding window is extremely compute +# expensive so for an exhaustive search do not expect the CNN to be real-time. + +import sensor, image, time, os, tf + +sensor.reset() # Reset and initialize the sensor. +sensor.set_pixformat(sensor.RGB565) # Set pixel format to RGB565 (or GRAYSCALE) +sensor.set_framesize(sensor.QVGA) # Set frame size to QVGA (320x240) +sensor.set_windowing((240, 240)) # Set 240x240 window. +sensor.skip_frames(time=2000) # Let the camera adjust. + +mobilenet_version = "1" # 1 +mobilenet_width = "0.5" # 1.0, 0.75, 0.50, 0.25 +mobilenet_resolution = "128" # 224, 192, 160, 128 + +mobilenet = "mobilenet_v%s_%s_%s_quant.tflite" % (mobilenet_version, mobilenet_width, mobilenet_resolution) +labels = [line.rstrip('\n') for line in open("mobilenet_labels.txt")] + +clock = time.clock() +while(True): + clock.tick() + + img = sensor.snapshot() + + # net.classify() will run the network on an roi in the image (or on the whole image if the roi is not + # specified). A classification score output vector will be generated for each location. At each scale the + # detection window is moved around in the ROI using x_overlap (0-1) and y_overlap (0-1) as a guide. + # If you set the overlap to 0.5 then each detection window will overlap the previous one by 50%. Note + # the computational work load goes WAY up the more overlap. Finally, for multi-scale matching after + # sliding the network around in the x/y dimensions the detection window will shrink by scale_mul (0-1) + # down to min_scale (0-1). For example, if scale_mul is 0.5 the detection window will shrink by 50%. + # Note that at a lower scale there's even more area to search if x_overlap and y_overlap are small... + + # default settings just do one detection... change them to search the image... + for obj in tf.classify(mobilenet, img, min_scale=1.0, scale_mul=0.5, x_overlap=0.0, y_overlap=0.0): + print("**********\nTop 5 Detections at [x=%d,y=%d,w=%d,h=%d]" % obj.rect()) + img.draw_rectangle(obj.rect()) + # This combines the labels and confidence values into a list of tuples + # and then sorts that list by the confidence values. + sorted_list = sorted(zip(labels, obj.output()), key = lambda x: x[1], reverse = True) + for i in range(5): + print("%s = %f" % (sorted_list[i][0], sorted_list[i][1])) + print(clock.fps(), "fps") diff --git a/scripts/examples/25-Machine-Learning/tf_mobilenet_serach_just_center.py b/scripts/examples/25-Machine-Learning/tf_mobilenet_serach_just_center.py index e69de29bb..1c243533c 100644 --- a/scripts/examples/25-Machine-Learning/tf_mobilenet_serach_just_center.py +++ b/scripts/examples/25-Machine-Learning/tf_mobilenet_serach_just_center.py @@ -0,0 +1,62 @@ +# TensorFlow Lite Mobilenet V1 Example +# +# Google's Mobilenet V1 detects 1000 classes of objects +# +# WARNING: Mobilenet is trained on ImageNet and isn't meant to classify anything +# in the real world. It's just designed to score well on the ImageNet dataset. +# This example just shows off running mobilenet on the OpenMV Cam. However, the +# default model is not really usable for anything. You have to use transfer +# learning to apply the model to a target problem by re-training the model. +# +# NOTE: This example only works on the OpenMV Cam H7 Pro (that has SDRAM) and better! +# +# In this example we slide the detector window over the image and get a list +# of activations. Note that use a CNN with a sliding window is extremely compute +# expensive so for an exhaustive search do not expect the CNN to be real-time. + +import sensor, image, time, os, tf + +sensor.reset() # Reset and initialize the sensor. +sensor.set_pixformat(sensor.RGB565) # Set pixel format to RGB565 (or GRAYSCALE) +sensor.set_framesize(sensor.QVGA) # Set frame size to QVGA (320x240) +sensor.set_windowing((240, 240)) # Set 240x240 window. +sensor.skip_frames(time=2000) # Let the camera adjust. + +mobilenet_version = "1" # 1 +mobilenet_width = "0.5" # 1.0, 0.75, 0.50, 0.25 +mobilenet_resolution = "128" # 224, 192, 160, 128 + +mobilenet = "mobilenet_v%s_%s_%s_quant.tflite" % (mobilenet_version, mobilenet_width, mobilenet_resolution) +labels = [line.rstrip('\n') for line in open("mobilenet_labels.txt")] + +clock = time.clock() +while(True): + clock.tick() + + img = sensor.snapshot() + + # net.classify() will run the network on an roi in the image (or on the whole image if the roi is not + # specified). A classification score output vector will be generated for each location. At each scale the + # detection window is moved around in the ROI using x_overlap (0-1) and y_overlap (0-1) as a guide. + # If you set the overlap to 0.5 then each detection window will overlap the previous one by 50%. Note + # the computational work load goes WAY up the more overlap. Finally, for multi-scale matching after + # sliding the network around in the x/y dimensions the detection window will shrink by scale_mul (0-1) + # down to min_scale (0-1). For example, if scale_mul is 0.5 the detection window will shrink by 50%. + # Note that at a lower scale there's even more area to search if x_overlap and y_overlap are small... + + # Setting x_overlap=-1 forces the window to stay centered in the ROI in the x direction always. If + # y_overlap is not -1 the method will search in all vertical positions. + + # Setting y_overlap=-1 forces the window to stay centered in the ROI in the y direction always. If + # x_overlap is not -1 the method will serach in all horizontal positions. + + # default settings just do one detection... change them to search the image... + for obj in tf.classify(mobilenet, img, min_scale=1.0, scale_mul=0.5, x_overlap=-1, y_overlap=-1): + print("**********\nTop 5 Detections at [x=%d,y=%d,w=%d,h=%d]" % obj.rect()) + img.draw_rectangle(obj.rect()) + # This combines the labels and confidence values into a list of tuples + # and then sorts that list by the confidence values. + sorted_list = sorted(zip(labels, obj.output()), key = lambda x: x[1], reverse = True) + for i in range(5): + print("%s = %f" % (sorted_list[i][0], sorted_list[i][1])) + print(clock.fps(), "fps")