deep-learning-with-pytorch/p2ch11/vis.py

import matplotlib
matplotlib.use('nbagg')

import numpy as np
import matplotlib.pyplot as plt

from p2ch11.dsets import Ct, LunaDataset

clim=(-1000.0, 300)

def findPositiveSamples(start_ndx=0, limit=100):
    ds = LunaDataset(sortby_str='label_and_size')

    positiveSample_list = []
    for sample_tup in ds.candidateInfo_list:
        if sample_tup.isNodule_bool:
            print(len(positiveSample_list), sample_tup)
            positiveSample_list.append(sample_tup)

        if len(positiveSample_list) >= limit:
            break

    return positiveSample_list

def showCandidate(series_uid, batch_ndx=None, **kwargs):
    ds = LunaDataset(series_uid=series_uid, **kwargs)
    pos_list = [i for i, x in enumerate(ds.candidateInfo_list) if x.isNodule_bool]

    if batch_ndx is None:
        if pos_list:
            batch_ndx = pos_list[0]
        else:
            print("Warning: no positive samples found; using first negative sample.")
            batch_ndx = 0

    ct = Ct(series_uid)
    ct_t, pos_t, series_uid, center_irc = ds[batch_ndx]
    ct_a = ct_t[0].numpy()

    fig = plt.figure(figsize=(30, 50))

    group_list = [
        [9, 11, 13],
        [15, 16, 17],
        [19, 21, 23],
    ]

    subplot = fig.add_subplot(len(group_list) + 2, 3, 1)
    subplot.set_title('index {}'.format(int(center_irc.index)), fontsize=30)
    for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
        label.set_fontsize(20)
    plt.imshow(ct.hu_a[int(center_irc.index)], clim=clim, cmap='gray')

    subplot = fig.add_subplot(len(group_list) + 2, 3, 2)
    subplot.set_title('row {}'.format(int(center_irc.row)), fontsize=30)
    for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
        label.set_fontsize(20)
    plt.imshow(ct.hu_a[:,int(center_irc.row)], clim=clim, cmap='gray')
    plt.gca().invert_yaxis()

    subplot = fig.add_subplot(len(group_list) + 2, 3, 3)
    subplot.set_title('col {}'.format(int(center_irc.col)), fontsize=30)
    for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
        label.set_fontsize(20)
    plt.imshow(ct.hu_a[:,:,int(center_irc.col)], clim=clim, cmap='gray')
    plt.gca().invert_yaxis()

    subplot = fig.add_subplot(len(group_list) + 2, 3, 4)
    subplot.set_title('index {}'.format(int(center_irc.index)), fontsize=30)
    for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
        label.set_fontsize(20)
    plt.imshow(ct_a[ct_a.shape[0]//2], clim=clim, cmap='gray')

    subplot = fig.add_subplot(len(group_list) + 2, 3, 5)
    subplot.set_title('row {}'.format(int(center_irc.row)), fontsize=30)
    for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
        label.set_fontsize(20)
    plt.imshow(ct_a[:,ct_a.shape[1]//2], clim=clim, cmap='gray')
    plt.gca().invert_yaxis()

    subplot = fig.add_subplot(len(group_list) + 2, 3, 6)
    subplot.set_title('col {}'.format(int(center_irc.col)), fontsize=30)
    for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
        label.set_fontsize(20)
    plt.imshow(ct_a[:,:,ct_a.shape[2]//2], clim=clim, cmap='gray')
    plt.gca().invert_yaxis()

    for row, index_list in enumerate(group_list):
        for col, index in enumerate(index_list):
            subplot = fig.add_subplot(len(group_list) + 2, 3, row * 3 + col + 7)
            subplot.set_title('slice {}'.format(index), fontsize=30)
            for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
                label.set_fontsize(20)
            plt.imshow(ct_a[index], clim=clim, cmap='gray')


    print(series_uid, batch_ndx, bool(pos_t[0]), pos_list)