o - iKM@sddlmZddlmZddlmZddlZddlmZddl mm Z ddl m Z ddlmZddlmZmZddlmZgd ZGd d d ejZGd d d ejZGdddejZGdddejZGdddejZGdddejZeZZ Z!dS)) annotations)Sequence)OptionalN) Convolution)UpSample)ConvPool)ensure_tuple_rep)BasicUnetPlusPlusBasicunetPlusPlusbasicunetplusplusBasicUNetPlusPlusKernelModifiedcs(eZdZdZfddZddZZS)Attention_blockz Attention Block c stt|ttj||dddddt||_ttj||dddddt||_ttj|ddddddtdt |_ tj dd|_ dS)NrT) kernel_sizestridepaddingbias)inplace) superr__init__nn SequentialConv3d BatchNorm3dW_gW_xSigmoidpsiReLUrelu)selfF_gF_lF_int __class__q/home/dell461/cl/sdc2/last_ska_mid/HISourceFinder-master-l/src/monai/networks/nets/basic_unetplusplus_modified.pyr(szAttention_block.__init__cCs4||}||}|||}||}||S)N)rrr r)r!gxg1x1rr'r'r(forward<s   zAttention_block.forward__name__ __module__ __qualname____doc__rr- __classcell__r'r'r%r(r#s rc@seZdZdZdddZdS) MCDropout3du>MC Dropout：无论 model.train()/eval() 都启用随机失活input torch.TensorreturncCstj||jd|jdS)NT)ptrainingr)F dropout3dr8r)r!r5r'r'r(r-GszMCDropout3d.forwardN)r5r6r7r6)r/r0r1r2r-r'r'r'r(r4Dsr4c,eZdZdZ    ddfdd ZZS)TwoConvztwo convolutions.r@r@rrr@r spatial_dimsintin_chnsout_chnsact str | tuplenormrbooldropout float | tupler Sequence[int]rOptional[tuple]rSequence[int] | intc s\tt||||||||| | d } t||||||||| d } |d| |d| dS)a Args: spatial_dims: number of spatial dimensions. in_chns: number of input channels. out_chns: number of output channels. act: activation type and arguments. norm: feature normalization type and arguments. bias: whether to have a bias term in convolution blocks. dropout: dropout ratio. Defaults to no dropout. )rFrHrJrrrstrides)rFrHrJrrrconv_0conv_1N)rrr add_module) r!rBrDrErFrHrrJrrrrPrQr%r'r(rNs4   zTwoConv.__init__)r>r?rAr)rBrCrDrCrErCrFrGrHrGrrIrJrKrrLrrMrrNr/r0r1r2rr3r'r'r%r(r=K r=cr<)Downz-maxpooling downsampling and two convolutions.r>r?rAmaxpoolrBrCrDrErFrGrHrrIrJrKrrLrrMdownsample_modestrc  s~t| dkrtd|fdd} |d| d} n| dkr"d} ntd| t||||||||| | d } |d | d S) a Args: spatial_dims: number of spatial dimensions. in_chns: number of input channels. out_chns: number of output channels. act: activation type and arguments. norm: feature normalization type and arguments. bias: whether to have a bias term in convolution blocks. dropout: dropout ratio. Defaults to no dropout. downsample_mode: 'maxpool' or 'strideconv'. Defaults to 'maxpool'. rVMAXr[rr max_poolingr strideconvzUnsupported downsample mode: )rrrconvsN)rrrrR ValueErrorr=)r!rBrDrErFrHrrJrrrWr]Z conv_strider_r%r'r(rs*   z Down.__init__)r>r?rArV)rBrCrDrCrErCrFrGrHrGrrIrJrKrrLrrMrWrXrSr'r'r%r(rUrTrUcsBeZdZdZ          d,d-fd$d% Zd.d*d+ZZS)/UpCatzHupsampling, concatenation with the encoder feature map, two convolutionsr>deconvdefaultlinearTr?rAFrBrCrDcat_chnsrErFrGrHrrIrJrKupsamplerXpre_convnn.Module | str | None interp_mode align_corners bool | Nonehalvesis_padrrLrrM attentionc st| dkr| dur|}n| r|dn|}t|||d| | | | |d |_t||||||||||d |_||_||_|jrKt|||dd|_ dSdS)a Args: spatial_dims: number of spatial dimensions. in_chns: number of input channels to be upsampled. cat_chns: number of channels from the encoder. out_chns: number of output channels. act: activation type and arguments. norm: feature normalization type and arguments. bias: whether to have a bias term in convolution blocks. dropout: dropout ratio. Defaults to no dropout. upsample: upsampling mode, available options are ``"deconv"``, ``"pixelshuffle"``, ``"nontrainable"``. pre_conv: a conv block applied before upsampling. Only used in the "nontrainable" or "pixelshuffle" mode. interp_mode: {``"nearest"``, ``"linear"``, ``"bilinear"``, ``"bicubic"``, ``"trilinear"``} Only used in the "nontrainable" mode. align_corners: set the align_corners parameter for upsample. Defaults to True. Only used in the "nontrainable" mode. halves: whether to halve the number of channels during upsampling. This parameter does not work on ``nontrainable`` mode if ``pre_conv`` is `None`. is_pad: whether to pad upsampling features to fit features from encoder. Defaults to True. attention: whether to use attention gate. Defaults to False. nontrainableNr[rZ)modergrirjr)rr)r"r#r$) rrrrfr=r_rm use_attentionrattention_gate)r!rBrDrerErFrHrrJrfrgrirjrlrmrrrnup_chnsr%r'r(rsB ,    zUpCat.__init__r*r6x_eOptional[torch.Tensor]cCs||}|dur\|}|jr|j||d}|jrNt|jd}dg|d}t|D]}|j| d|j| dkrDd||dd<q*tjj ||d}| tj ||gdd}|S| |}|S)z Args: x: features to be upsampled. x_e: optional features from the encoder, if None, this branch is not in use. N)r)r*r[rr replicatedim) rfrqrrrmlenshaperangetorchr functionalpadr_cat)r!r*rtx_0Z x_e_non_none dimensionsspir'r'r(r- s&    z UpCat.forward) r>rbrcrdTTTr?rAF)"rBrCrDrCrerCrErCrFrGrHrGrrIrJrKrfrXrgrhrirXrjrkrlrIrmrIrrLrrMrnrI)r*r6rtrur.r'r'r%r(ras Nrac sReZdZddddddddd fd d difdd d d f d&fd d! Zd'd$d%ZZS)(r r@rr[) r@rF LeakyReLUg?T)negative_sloperinstanceaffiner>rbrBrC in_channels out_channelsfeaturesrLdeep_supervisionrIrFrGrHrrJrKrfrX dropout_pfloatc sZt||_| |_t|d} td| dt||| d|||| dd|_t|| d| d|||| dd|_ t|| d| d|||| dd|_ t|| d| d |||| dd d |_ t|| d | d |||| dd d |_ t || d| d| d|||| | d ddd |_t || d| d| d|||| | d ddd |_t || d | d| d|||| | d ddd |_t || d | d | d |||| | d ddd |_t || d| dd| d|||| | d ddd |_t || d| dd| d|||| | d ddd |_t || d | dd| d|||| | d ddd |_t || d| dd | d|||| | d ddd |_t || d| dd | d|||| | d ddd |_t || d| dd | d|||| | d ddd |_|jrn|jdkrnt|jnt|_td|f| d|dd|_td|f| d|dd|_td|f| d|dd|_td|f| d|dd|_ dS)a A UNet++ implementation with 1D/2D/3D supports. Based on: Zhou et al. "UNet++: A Nested U-Net Architecture for Medical Image Segmentation". 4th Deep Learning in Medical Image Analysis (DLMIA) Workshop, DOI: https://doi.org/10.48550/arXiv.1807.10165 Args: spatial_dims: number of spatial dimensions. Defaults to 3 for spatial 3D inputs. in_channels: number of input channels. Defaults to 1. out_channels: number of output channels. Defaults to 2. features: six integers as numbers of features. Defaults to ``(32, 32, 64, 128, 256, 32)``, - the first five values correspond to the five-level encoder feature sizes. - the last value corresponds to the feature size after the last upsampling. deep_supervision: whether to prune the network at inference time. Defaults to False. If true, returns a list, whose elements correspond to outputs at different nodes. act: activation type and arguments. Defaults to LeakyReLU. norm: feature normalization type and arguments. Defaults to instance norm. bias: whether to have a bias term in convolution blocks. Defaults to True. According to `Performance Tuning Guide `_, if a conv layer is directly followed by a batch norm layer, bias should be False. dropout: dropout ratio. Defaults to no dropout. upsample: upsampling mode, available options are ``"deconv"``, ``"pixelshuffle"``, ``"nontrainable"``. Examples:: # for spatial 2D >>> net = BasicUNetPlusPlus(spatial_dims=2, features=(64, 128, 256, 512, 1024, 128)) # for spatial 2D, with deep supervision enabled >>> net = BasicUNetPlusPlus(spatial_dims=2, features=(64, 128, 256, 512, 1024, 128), deep_supervision=True) # for spatial 2D, with group norm >>> net = BasicUNetPlusPlus(spatial_dims=2, features=(64, 128, 256, 512, 1024, 128), norm=("group", {"num_groups": 4})) # for spatial 3D >>> net = BasicUNetPlusPlus(spatial_dims=3, features=(32, 32, 64, 128, 256, 32)) See Also - :py:class:`monai.networks.nets.BasicUNet` - :py:class:`monai.networks.nets.DynUNet` - :py:class:`monai.networks.nets.UNet` zBasicUNetPlusPlus features: .r)r@r@r\rr[r@r^)rrWFT)rlrrnconvN)!rrrrr printr=conv_0_0rUconv_1_0conv_2_0conv_3_0conv_4_0ra upcat_0_1 upcat_1_1 upcat_2_1 upcat_3_1 upcat_0_2 upcat_1_2 upcat_2_2 upcat_0_3 upcat_1_3 upcat_0_4r4rIdentitymc_dropout_outrfinal_conv_0_1final_conv_0_2final_conv_0_3final_conv_0_4) r!rBrrrrrFrHrrJrfrfear%r'r(r)s A                       z(BasicUNetPlusPlusKernelModified.__init__r*r6cCs$||}||}|||}||}|||}||tj||gdd}||}| ||} | | tj||gdd} | | tj|||gdd} | |} | | |} || tj|| gdd}||tj||| gdd}||tj|||| gdd}||}||}|g}|S)a Args: x: input should have spatially N dimensions ``(Batch, in_channels, dim_0[, dim_1, ..., dim_N-1])``, N is defined by `dimensions`. It is recommended to have ``dim_n % 16 == 0`` to ensure all maxpooling inputs have even edge lengths. Returns: A torch Tensor of "raw" predictions in shape ``(Batch, out_channels, dim_0[, dim_1, ..., dim_N-1])``. rrw)rrrrrrr|rrrrrrrrrrrr)r!r*x_0_0x_1_0x_0_1x_2_0x_1_1x_0_2x_3_0x_2_1x_1_2x_0_3x_4_0x_3_1x_2_2x_1_3x_0_4 output_0_4outputr'r'r(r-Gs&          z'BasicUNetPlusPlusKernelModified.forward)rBrCrrCrrCrrLrrIrFrGrHrGrrIrJrKrfrXrr)r*r6)r/r0r1rr-r3r'r'r%r(r (s    r )" __future__rcollections.abcrtypingrr|torch.nnrtorch.nn.functionalr}r:"monai.networks.blocks.convolutionsrmonai.networks.blocks.upsamplermonai.networks.layers.factoriesrrmonai.utils.miscr __all__Moduler Dropout3dr4rr=rUrar r r r r'r'r'r(s*      !86oO