Created TensorFlow text graphs (markdown)

TensorFlow-text-graphs.md

@@ -0,0 +1,203 @@
+To read networks from TensorFlow framework there is [cv::dnn::readNetFromTensorflow](https://docs.opencv.org/master/d6/d0f/group__dnn.html#gad820b280978d06773234ba6841e77e8d) method which can work with `.pb` files with frozen TensorFlow graph. However sometimes it is not enough to have only `.pb` file to import network into OpenCV.
+
+Depends on topology, graph may contains some unfused layers which are not covered by internal subgraphs fusion procedure. In example, instead of a single `Flatten` node for tensors flattening (reshaping from 4D `[N, C, H, W]` to 2D `[N, C*H*W]`) graph will have a set of nodes that precompute final shape and then call `Reshape` operation:
+
+![image](https://user-images.githubusercontent.com/25801568/74974976-3fe30b00-5437-11ea-8512-dc9d66c497ea.png)
+
+If you know how to replace TensorFlow subgraph to something simpler which can be parser by OpenCV (see [tf_importer.cpp](https://github.com/opencv/opencv/blob/master/modules/dnn/src/tensorflow/tf_importer.cpp) and [tf_graph_simplifier.cpp](https://github.com/opencv/opencv/blob/master/modules/dnn/src/tensorflow/tf_graph_simplifier.cpp)) you may use **optional** parameter of [cv::dnn::readNetFromTensorflow](https://docs.opencv.org/master/d6/d0f/group__dnn.html#gad820b280978d06773234ba6841e77e8d) to get a text graph representation of a frozen `.pb` file.
+
+Whet is text graph representation? TensorFlow uses Google Protobuf format to store deep learning networks. This format can be as binary as textual. Due models usually have a lot of trainable weights it is more optimal to store them as binary rather than text file.
+
+To generate `.pbtxt` file for OpenCV, you need to convert a `.pb` model to text file. The following script can do it:
+
+```python
+import tensorflow as tf
+
+pb_file = 'model.pb'
+graph_def = tf.compat.v1.GraphDef()
+
+try:
+    with tf.io.gfile.GFile(pb_file, 'rb') as f:
+        graph_def.ParseFromString(f.read())
+except:
+    with tf.gfile.FastGFile(pb_file, 'rb') as f:
+        graph_def.ParseFromString(f.read())
+
+# Delete weights
+for i in reversed(range(len(graph_def.node))):
+    if graph_def.node[i].op == 'Const':
+        del graph_def.node[i]
+
+graph_def.library.Clear()
+
+tf.compat.v1.train.write_graph(graph_def, "", 'model.pbtxt', as_text=True)
+```
+
+Then replace a problematic subgraph
+```yaml
+node {
+  name: "flatten_1/Shape"
+  op: "Shape"
+  input: "max_pooling2d_2/MaxPool"
+  attr {
+    key: "T"
+    value {
+      type: DT_FLOAT
+    }
+  }
+  attr {
+    key: "out_type"
+    value {
+      type: DT_INT32
+    }
+  }
+}
+node {
+  name: "flatten_1/strided_slice"
+  op: "StridedSlice"
+  input: "flatten_1/Shape"
+  input: "flatten_1/strided_slice/stack"
+  input: "flatten_1/strided_slice/stack_1"
+  input: "flatten_1/strided_slice/stack_2"
+  attr {
+    key: "Index"
+    value {
+      type: DT_INT32
+    }
+  }
+  attr {
+    key: "T"
+    value {
+      type: DT_INT32
+    }
+  }
+  attr {
+    key: "begin_mask"
+    value {
+      i: 0
+    }
+  }
+  attr {
+    key: "ellipsis_mask"
+    value {
+      i: 0
+    }
+  }
+  attr {
+    key: "end_mask"
+    value {
+      i: 1
+    }
+  }
+  attr {
+    key: "new_axis_mask"
+    value {
+      i: 0
+    }
+  }
+  attr {
+    key: "shrink_axis_mask"
+    value {
+      i: 0
+    }
+  }
+}
+node {
+  name: "flatten_1/Prod"
+  op: "Prod"
+  input: "flatten_1/strided_slice"
+  input: "flatten_1/Const"
+  attr {
+    key: "T"
+    value {
+      type: DT_INT32
+    }
+  }
+  attr {
+    key: "Tidx"
+    value {
+      type: DT_INT32
+    }
+  }
+  attr {
+    key: "keep_dims"
+    value {
+      b: false
+    }
+  }
+}
+node {
+  name: "flatten_1/stack_3862"
+  op: "Pack"
+  input: "flatten_1/stack_3862/0"
+  input: "flatten_1/Prod"
+  attr {
+    key: "N"
+    value {
+      i: 2
+    }
+  }
+  attr {
+    key: "T"
+    value {
+      type: DT_INT32
+    }
+  }
+  attr {
+    key: "axis"
+    value {
+      i: 0
+    }
+  }
+}
+node {
+  name: "flatten_1/Reshape"
+  op: "Reshape"
+  input: "max_pooling2d_2/MaxPool"
+  input: "flatten_1/stack_3862"
+  attr {
+    key: "T"
+    value {
+      type: DT_FLOAT
+    }
+  }
+  attr {
+    key: "Tshape"
+    value {
+      type: DT_INT32
+    }
+  }
+}
+```
+
+to a single `Flatten` layer which can be easily imported by OpenCV:
+
+```yaml
+node {
+  name: "flatten_1/Reshape"
+  op: "Flatten"
+  input: "max_pooling2d_2/MaxPool"
+}
+```
+
+**NOTE**: check that after removing the nodes connections between nodes are not broken. That's why we use the same name for new `Flatten` node because there is a consumer which is connected to this specific name:
+```yaml
+node {
+  name: "dense_1/MatMul"
+  op: "MatMul"
+  input: "flatten_1/Reshape"
+  input: "dense_1/MatMul/ReadVariableOp"
+  attr {
+    key: "T"
+    value { type: DT_FLOAT }
+  }
+  attr {
+    key: "transpose_a"
+    value { b: false }
+  }
+  attr {
+    key: "transpose_b"
+    value { b: false }
+  }
+}
+```