Functions — github.com/TJUMMG/3DFP_FCGAN

CodeHub MCPcopy Create free account

hub / github.com/TJUMMG/3DFP_FCGAN / functions

Functions67 in github.com/TJUMMG/3DFP_FCGAN

⨍ Functions 67 ◇ Types & classes 3

↓ 46 callersFunctiongen_conv

Define conv for generator. Args: x: Input. cnum: Channel number. ksize: Kernel size. Stride: Convolution stride.

inpaint_ops.py:19

↓ 11 callersFunctionconv2d

(inputs, name, out_channels, bn = False, is_training = False ,activation=False, ksize=3, stride = 1)

util/util.py:17

↓ 10 callersFunctionresBlock

(x, num_outputs, kernel_size = 4, stride=1, activation_fn=tf.nn.relu, normalizer_fn=tcl.batch_norm, scope=None

↓ 8 callersFunctiondis_conv

Define conv for discriminator. Activation is set to leaky_relu. Args: x: Input. cnum: Channel number. ksize: Kernel s

inpaint_ops.py:74

↓ 7 callersFunctionconv2d

(inputs, name, out_channels, bn = False, is_training = False ,activation=False, ksize=3, stride = 1)

util/util_bn.py:17

↓ 6 callersFunctionlocal_patch

Crop local patch according to bbox. Args: x: input bbox: (top, left, height, width) Returns: tf.Tensor: local patch

inpaint_ops.py:152

↓ 5 callersFunctionconv

(x, channels, kernel=4, stride=2, pad=0, pad_type='zero', use_bias=True, sn=False, scope='saconv_0')

↓ 4 callersFunctionconv_block

(x,output_num,is_training,name)

util/util_bn.py:72

↓ 4 callersFunctionconv_block

(x,output_num,is_training,name)

util/util.py:66

↓ 4 callersFunctiongen_deconv

Define deconv for generator. The deconv is defined to be a x2 resize_nearest_neighbor operation with additional gen_conv operation. Args:

inpaint_ops.py:50

↓ 4 callersFunctionspectral_norm

(w, iteration=1)

↓ 3 callersMethodbuild_inpaint_net

Inpaint network. Args: x: incomplete image, [-1, 1] mask: mask region {0, 1} Returns: [-1, 1] as

↓ 3 callersFunctionflatten

↓ 3 callersFunctionhw_flatten

↓ 3 callersFunctionrelu

↓ 2 callersFunctionbatch_norm

(x, is_training=True, scope='sabatch_norm')

↓ 2 callersFunctionbbox2mask

Generate mask tensor from bbox. Args: bbox: configuration tuple, (top, left, height, width) config: Config should have configurat

inpaint_ops.py:120

↓ 2 callersMethodbuild_wgan_discriminator

(self, batch_local, batch_global, reuse=False, training=True)

↓ 2 callersFunctioncontextual_attention

Contextual attention layer implementation. Contextual attention is first introduced in publication: Generative Image Inpainting with Con

inpaint_ops.py:216

↓ 2 callersMethodfind_eye_nose_mouth

(self, pos_map,config)

↓ 2 callersFunctionhour_glass

(x, output_num, depth, is_training, name)

util/util.py:88

↓ 2 callersFunctionl2_norm

↓ 2 callersFunctionmake_color_wheel

inpaint_ops.py:387

↓ 2 callersFunctionrandom_bbox

Generate a random tlhw with configuration. Args: config: Config should have configuration including IMG_SHAPES, VERTICAL_MARG

inpaint_ops.py:95

↓ 2 callersFunctionspatial_discounting_mask

Generate spatial discounting mask constant. Spatial discounting mask is first introduced in publication: Generative Image Inpainting with

inpaint_ops.py:184

↓ 1 callersMethodbuild_infer_graph

↓ 1 callersMethodbuild_server_graph

↓ 1 callersMethodbuild_wgan_global_discriminator

(self, x, reuse=False, training=True)

↓ 1 callersMethodbuild_wgan_local_discriminator

(self, x, reuse=False, training=True)

↓ 1 callersFunctioncompute_color

inpaint_ops.py:421

↓ 1 callersFunctionflow_to_image_tf

Tensorflow ops for computing flow to image.

inpaint_ops.py:480

↓ 1 callersFunctionmain

cropped_image.py:12

↓ 1 callersFunctionresblock

(x_init, channels, use_bias=True, is_training=True, sn=False, scope='saresblock')

↓ 1 callersFunctionresize_mask_like

Resize mask like shape of x. Args: mask: Original mask. x: To shape of x. Returns: tf.Tensor: resized mask

inpaint_ops.py:167

↓ 1 callersFunctiontest_contextual_attention

Test contextual attention layer with 3-channel image input (instead of n-channel feature).

inpaint_ops.py:349

(self, x, is_training = True)

predictor.py:30

(self, resolution_inp = 256, resolution_op = 256, channel = 3, name = 'resfcn256')

predictor.py:24

(self, resolution_inp = 256, resolution_op = 256)

predictor.py:80

Functionattention

(x, ch, sn=False, scope = 'attention',reuse=False)

inpaint_ops.py:329

Methodbuild_graph_with_losses

(self, batch_data, config, training=True, summary=False, reuse=False)

Methodbuild_static_infer_graph

(x, channels, kernel=4, stride=2, padding='SAME', use_bias=True, sn=False, scope='sadeconv_0')

Functiondeconv2d

(inputs, name, out_channels, bn = False, is_training = False ,activation=False, ksize = 4, stride = 2)

util/util_bn.py:35

Functiondeconv2d

(inputs, name, out_channels, bn = False, is_training = False ,activation=False, ksize = 4, stride = 2)

util/util.py:33

Functiondiscriminator_loss

(loss_func, real, fake)

Functionflow_to_image

Transfer flow map to image. Part of code forked from flownet.

inpaint_ops.py:451

Functionfully_conneted

(x, units, use_bias=True, sn=False, scope='safully_0')

Functiongenerator_loss

(loss_func, fake)

Functionglobal_avg_pooling

Functionhighlight_flow

Convert flow into middlebury color code image.

inpaint_ops.py:490

Functionhighlight_flow_tf

Tensorflow ops for highlight flow.

inpaint_ops.py:508

Functionhour_glass

(x, output_num, depth, is_training, name)

util/util_bn.py:98

Functionim2uint8

util/util_bn.py:9

Functionim2uint8

Functionimage2edge

Convert image to edges.

inpaint_ops.py:518

(bbox, height, width, delta_h, delta_w)

inpaint_ops.py:132

predictor.py:100

Methodpredict_batch

predictor.py:109

Functionresblock

(inputs,out_channels,is_training,name)

util/util_bn.py:57

Functionresblock

(inputs,out_channels,is_training,name)

util/util.py:53

(self, model_path)

predictor.py:97

Functionstr2bool

inpaint_ops.py:326

Functionup_sample

(x, scale_factor=2)

predictor.py:73