Single　image　super-resolution　(SISR)　is　an　emerging　application　in　medical　imaging,　as　high-quality　images　need　to　be　obtained　with　limited　radiation　dose,　such　as　low-dose　computed　tomography　and　low-field　magnetic　resonance　imaging.　However,　a　certain　amount　of　noise　and　artifacts　are　frequently　present　in　medical　images　due　to　the　constraints　of　imaging　equipment　and　the　surrounding　environment.　This　can　cause　structural　distortion　and　blurring　of　details　in　the　resulting　medical　images.　This　research　proposes　a　dual-domain　residual　convolutional　neural　network　(DDRN)　based　super-resolution　reconstruction　technique　for　medical　images.　Firstly,　shallow　feature　extraction　is　performed　on　the　low-resolution　images　through　convolutional　networks.　Secondly,　a　newly　designed　spatial　domain　residual　block　(SDRB)　is　employed　to　alleviate　gradient　vanishing　issues　while　enhancing　feature　reuse,　thereby　facilitating　the　recovery　of　edge　details.　Additionally,　a　coordinate　attention　(CA)　module　is　incorporated　to　capture　both　channel-wise　and　long-range　spatial　correlations.　By　assigning　adaptive　weights　to　different　channels　and　capturing　global　spatial　context,　CA　enables　precise　restoration　of　texture　and　structural　details　in　medical　images.　Subsequently,　parallel　wavelet　domain　residual　blocks　(WDRB)　are　employed　to　capture　multi-directional　high-frequency　information,　facilitating　the　restoration　of　clear　texture　details.　Lastly,　by　introducing　gradient　space　loss　for　training　guidance,　the　network　is　encouraged　to　focus　more　on　restoring　the　geometric　structure　of　the　image　and　suppressing　artifacts.　Extensive　experiments　demonstrate　the　superior　performance　of　DDRN.　At　a　scaling　factor　of　4,　the　network　achieves　a　peak　signal-to-noise　ratio　(PSNR)　of　37.28　dB　and　a　structural　similarity　index　(SSIM)　of　0.9310　on　the　COVID-19　CT　lung　segmentation　dataset,　and　a　PSNR　of　28.85　dB　and　an　SSIM　of　0.9012　on　the　MRBrain2018　dataset.