Functional Sample

app.py

@@ -6,6 +6,8 @@ import numpy as np
 
 from huggingface_hub import from_pretrained_keras
 
+from einops import repeat
+
 custom_objects = {'BilinearUpSampling2D': BilinearUpSampling2D, 'depth_loss_function': None}
 print('Loading model...')
 model = from_pretrained_keras("keras-io/monocular-depth-estimation", custom_objects=custom_objects, compile=False)
@@ -15,56 +17,82 @@ import importlib
 import utils
 importlib.reload(utils)
 
+def layer_over_image(raw_image, filter, custom_color = [0, 0, 0]):
+    # print(raw_image[:, :, 0])
+    out_image = raw_image
+    out_image[:,:,0] = raw_image[:, :, 0] * filter
+    out_image[:,:,1] = raw_image[:, :, 1] * filter
+    out_image[:,:,2] = raw_image[:, :, 2] * filter
+    return raw_image
+
 def infer(image, min_th, max_th):
-    print('_'*20)
+    # print('_'*20)
     inputs = utils.load_images([image])
     outputs = utils.predict(model, inputs)
 
     plasma = plt.get_cmap('plasma')
     rescaled = outputs[0][:, :, 0]
-    print("Min Max Bef", np.min(rescaled), np.max(rescaled))
+    # print("Min Max Bef", np.min(rescaled), np.max(rescaled))
     rescaled = rescaled - np.min(rescaled)
     rescaled = rescaled / np.max(rescaled)
 
-    image_out = plasma(rescaled)[:, :, :3]
+    im_heat = plasma(rescaled)[:, :, :3]
+
+    # print("Min Max Aft", np.min(rescaled), np.max(rescaled))
+
+    # print("Shape Scaled:",rescaled.shape)
+    filt_base = rescaled
+    filt_base = repeat(filt_base, "h w -> (h 2) (w 2)")
+    filt_arr_min = (filt_base > min_th/100)
+    filt_arr_max = (filt_base < max_th/100)
+
+    filt_arr = filt_arr_min * filt_base * filt_arr_max
+    im_heat_filt = plasma(filt_arr)[:, :, :3]
 
-    print("Min Max Aft", np.min(rescaled), np.max(rescaled))
+    if max_th < 100:
+        image_emph = layer_over_image(image, filt_arr_max)
+    else:
+        image_emph = image
+    
+    if min_th > 0:
+        image_emph = layer_over_image(image, filt_arr_min)
+    else:
+        image_emph = image
 
-    print("Shape Scaled:",rescaled.shape)
-    filtered = rescaled
-    # filtered[filtered[:, :, 0] < min_th/100, 0] = 0
-    # filtered[filtered[:, :, 0] < min_th/100, 1] = 0
-    # filtered[filtered[:, :, 0] < min_th/100, 2] = 0
-    # filt_arr = filtered[((filtered[:,0] > min_th/100) & (filtered[:,0] < max_th/100))]
-    filt_arr = (filtered > min_th/100) * filtered * (filtered < max_th/100)
 
+    # print("filt arr min", filt_arr_min)
 
-    print("Shape Image:",image.shape)
-    print("Shape Image filt:",im_filt.shape)
-    print("Shape Image Heat:",image_out.shape)
-    im_filt = plasma(filt_arr)[:, :, :3]
-    return image_out, im_filt, image
+    # print("Shape Image:",image.shape)
+    # print("Shape Image filt:",im_heat_filt.shape)
+    # print("Shape Image Heat:",im_heat.shape)
+    return im_heat, image_emph
 
 # def detr(im):
 #     return im
 
 gr_input = [
-    gr.inputs.Image(label="image", type="numpy", shape=(640, 480))
-    ,gr.inputs.Slider(minimum=0, maximum=100, step=5, default=0, label="Minimum Threshold")
-    ,gr.inputs.Slider(minimum=0, maximum=100, step=5, default=100, label="Maximum Threshold")
+    gr.inputs.Image(label="Image", type="numpy", shape=(640, 480))
+    ,gr.inputs.Slider(minimum=0, maximum=100, step=0.5, default=0, label="Minimum Threshold")
+    ,gr.inputs.Slider(minimum=0, maximum=100, step=0.5, default=100, label="Maximum Threshold")
 ]
 
 gr_output = [
-    gr.outputs.Image(type="pil",label="HeatMap Image"),
-    gr.outputs.Image(type="pil",label="Filtered Image"),
-    gr.outputs.Image(type="pil",label="Output Image")
+    gr.outputs.Image(type="pil",label="Depth HeatMap"),
+    # gr.outputs.Image(type="pil",label="Filtered Image"),
+    # gr.outputs.Image(type="pil",label="Before"),
+    gr.outputs.Image(type="pil",label="Important Areas")
 ]
 
 iface = gr.Interface(
     fn=infer,
-    title="Space Title Here",
-    description = "Description Here",
+    title="Monocular Depth Filter",
+    description = "Used Keras Depth Estimation Model for estimating the depth of areas in an image. Image is then filtered out to only show the selected areas",
     inputs  = gr_input,
-    outputs = gr_output
+    outputs = gr_output,
+    examples=[
+        ["examples/00015_colors.png", 0, 90]
+        ,["examples/car.jpg", 0, 30]
+        ,["examples/dash.jpg", 10, 55]
+        ]
     )
 iface.launch()

utils.py

@@ -12,8 +12,8 @@ def predict(model, images, minDepth=10, maxDepth=1000, batch_size=2):
     # Compute predictions
     predictions = model.predict(images, batch_size=batch_size)
     # Put in expected range
-    print("Max Depth:", np.amax(predictions), maxDepth)
-    print("Min Depth:", np.amin(predictions), minDepth)
+    # print("Max Depth:", np.amax(predictions), maxDepth)
+    # print("Min Depth:", np.amin(predictions), minDepth)
     return np.clip(depth_norm(predictions, maxDepth=maxDepth), minDepth, maxDepth) / maxDepth